The Galaxy Note 9 isn't exactly the most exciting new device. 

It has a near-identical design to last year's Galaxy Note 8 and it runs Android 8.1 Oreo, which isn't the latest version of Android. 

Just because it's not a revolutionary new device, however, doesn't mean Samsung didn't sneak a few cool features into its new $1,000 smartphone. 

Below are four of the most innovative and exciting aspects of the Galaxy Note 9. 

SEE ALSO: I've spent the last 24 hours with Samsung's new Galaxy Note 9 — here are my first impressions

NOW READ: How Samsung's new Galaxy Note 9 compares with last year's Galaxy Note 8

1. The S Pen

Samsung changed the color of the S Pen that comes with the blue Galaxy Note 9 — rather than matching it to the phone's finish, the new S Pen is a contrasting bright yellow (with any of the three other colors in which you can get the device, the S Pen comes in a matching color). 

But beyond changing the color, Samsung upgraded the S Pen's insides, too.

The new S Pen has Bluetooth Low Energy, which means it can work with your phone up to 30 feet away, allowing you to use the S Pen as a remote —  click the button on the stylus to take a photo, change a presentation slide, or pause and play music and videos.

If there's one downside, it's that the new S Pen will need to charge from time to time. Thankfully, it does that while clipped in to the Galaxy Note 9, and Samsung says you'll get 30 more minutes of use after 40 seconds of charging. 

2. Dual aperture camera and 2x optical zoom

There are two of the major upgrades to the Galaxy Note 9's camera (which can be found on the Galaxy s9 as well):

• 2x optical zoom allows for significantly sharper zoomed-in images. Even phones like the Google Pixel 2, which has one of the best smartphone cameras in the world, doesn't have 2x optical zoom, and it shows in this camera comparison by my colleague Tony.
• Dual aperture may sound gimmicky, but it actually works. When you don't have a lot of light or your subject itself is dark, the Galaxy Note 9 can use its variable aperture to allow in more light. This results in better low-light photos — even better than the Pixel 2.

3. The intelligent camera

Samsung added a few new artificial intelligence features to its camera as well: 

• Scene optimizer recognizes what kind of object you're photographing and adjusts the settings automatically to take the best shot.
• Flaw detection will let you know if the image is blurry, if the subject blinked, if there is a smudge on the lens, or if there is backlight affecting the image quality.

• Researchers at CB Insights waded into Google's earnings, public statements and patent filings to illustrate where the search giant is headed.
• When it comes to artificial intelligence, Google is filing for more AI-related patents than Amazon, Microsoft and Facebook. 
• Their extensive report shows that when it comes to the amount of money Google commits to  overall R&D when compared to competitors, the search giant spends on the high-end.
• Google CEO Sundar Pichai said he believes AI will have more impact on our lives than the invention of fire. 

Researchers at CB Insights have released an extensive report on Google's business, illustrating just how dedicated the search company is to artificial intelligence.

CB Insights found that Google typically files each year more AI-related patents than the company's top rivals, including Amazon, Microsoft and Facebook.

Google's top execs have long said that AI is  core to the company's future growth, but CB Insights looked at a wide variety of data to try and learn how determined Google is to dominate AI. One area within AI that is of particular interest to Google is neural networks.

"Google is also focused on building out its deep learning capabilities," CB Insights researchers wrote, "which is more complex than traditional machine learning in that it generates predictions using an artificial neural network inspired by the human brain." 

CB Insights searched the patent filings and noted that among Google's top patent keywords, "neural network" was becoming more prominent.  

When it comes to overall Research & Development, the amount Google spends compared to rivals is on the high side, according to CB Insights--both in terms of absolute dollars and as a percentage of sales.

According to CB Insights, Google's plan is to defend the core search business while disrupting such industries as  transportation, logistics, and healthcare. But at the center of everything is AI.

"Unifying Alphabet’s approach across initiatives is its expertise in AI and machine learning," the researchers wrote, "which the company believes will help it become an all-encompassing service for both consumers and enterprises."

For Google, AI is a major strategic initiative. Even as Google uses AI to bolster its consumer services, its Google Cloud unit is working to gain ground on the cloud market-leading Amazon Web Services, and the second-place Microsoft Azure. 

And besides the competitive impetus, Google believes that AI is really the next big thing.

"AI is one of the most important things humanity is working on,"Google CEO Sundar Pichai said earlier this year. "It is more profound than, I don't know, electricity or fire."

SEE ALSO: A former Google China exec says the company's plan to build a censored search engine in the country is likely a violation of human rights

Join the conversation about this story »

NOW WATCH: How movie theaters are ruining your movie

• Investors and entrepreneurs say that the US may soon fall behind in the race to build better artificial intelligence.
• While many agree that the US was a leader in the technology's development, in recent years, the tables have turned: many say that China's tech could surpass the US's in five years or less. 

Artificial Intelligence is a complex and evolving technology. And whichever country masters it first will own "the black box of the future." 

That's how Ajeet Singh, founder of analytics software company ThoughtSpot, thinks about AI, which is expected to  play an increasingly important role in everyday life in the coming years.

"AI is the world's next big inflection point," says Singh. "It will hugely impact human productivity, the creation of drugs, the future of education and medicine."

If you're ahead on AI, he says, "that naturally means that you'll be ahead as a country."

For Singh, a serial entrepreneur who has built two data-based companies (one of which, Nutanix, is now valued at $9  billion on the public market), it's integral that the US start paying closer attention to the development of AI in China, where he says the technology is poised to take off.

"The future will be won or lost on this technology," said Singh. "I'm very concerned about the US falling behind. It's quite scary."

10 years ago, it was the other way around

While the US established itself as the uncontested technological leader in the internet's early years, many investors and entrepreneurs believe it could soon lose its top spot. 

Ron Cao, a partner at Shanghai-based firm Sky9 Capital, says China has increasingly become the place to look for new trends and innovations in the internet business. 

"What's going on with China's internet is a prediction of what will happen in the US internet market," says Cao. "10 years ago, it was completely the other way around: China was watching the US for cues. But over the past few years it's shifted. What's happening in China is beginning to be where it's going in the US."

While many experts agree that China still has yet to surpass the US in its development of AI (in April, an Oxford University researcher told Axios that he thought that China's AI was about half as good as that in the US), some say this won't be the case for long. 

AI is the next "space race"

"In the field of AI, China is neck and neck with the US," said Dennis Barrier, CEO of global venture firm Cathay Innovation. "I wouldn't say that China is ahead just yet. But it's very true that only a few years ago it was very far behind. You can see if the trend continues like that, they'll soon be ahead."

Some industry experts reckon that China's AI could outperform that of the US in less than a decade. Barrier, however, believes it will happen much sooner: Given China's staggering number of engineers, their indisputably aggressive work ethic, and the country's renewed interest in AI, "my personal feeling is that in two years, China will be ahead of the US in AI," said Barrier. "They're set to outpace us in the core tech that is transforming every field in the world."

While others have compared the race for better AI to the Cold War's arms race, Singh likens it to another global rivalry in innovation: "You can think of this in terms of the Space Race," he said. "The Space Race uplifted the industrial skillset in the country. I think this is quite similar in regards to staying ahead — we’ve got to make it a national priority."

SEE ALSO: Here's the latest hint that Apple will launch new Apple Watch models next month

Join the conversation about this story »

NOW WATCH: This machine turns a truck into a tent in seconds - here's how it's installed

• Facebook is trying to use AI to make MRI scans up to 10x faster.
• The social network's AI lab has teamed up with the NYU School of Medicine for a new research project.
• The company has access to 3 million magnetic resonance images from 10,000 clinical cases.

Facebook wants to revolutionise MRI (magnetic resonance imaging) scans.

The social networking giant announced on Monday a new research project that aims to use artificial intelligence (AI) to make MRI scans up to ten times faster, and has been granted access to a trove of around 3 million anonymised MRI images from 10,000 clinical cases.

In an announcement blog post, Facebook said the data it has accessed has had patient names removed, as well as "all other protected health information," and the project is compliant with HIPAA privacy regulations. 

The project is a collaboration between Facebook's FAIR AI research lab and NYU School of Medicine's Department of Radiology.

The efforts are a major foray by Facebook into into the medical space, as it attempts to apply its experimental AI research efforts to real-world problems. It's an area that is rich with possibilities, but fraught with potential problems and privacy concerns. DeepMind, an AI lab owned by Google faced years of scrutiny over a data-sharing partnership with Britain's NHS (National Health Service), and the UK's data regulator ultimately ruled that the organisation violated British privacy laws.

MRI scans are lengthy affairs, often taking half an hour or so, while the patient lies dead still in a cramped tube. With this project, Facebook is betting that time can be cut down radically by using AI to capture less data, focusing the scanning process only on what's important.

There are potential risks to the plan — what if in collecting less data, something is missed? — but also significant potential benefits. It means patients can take the scans more easily, and it increases the rate at which MRIs can be conducted, potentially bringing patient costs down.

"Using AI, it may be possible to capture less data and therefore scan faster, while preserving or even enhancing the rich information content of magnetic resonance images. The key is to train artificial neural networks to recognize the underlying structure of the images in order to fill in views omitted from the accelerated scan," Facebook wrote in the blog post.

"This approach is similar to how humans process sensory information. When we experience the world, our brains often receive an incomplete picture — as in the case of obscured or dimly lit objects — that we need to turn into actionable information."

Facebook has faced significant outrage over its privacy policies and use of customer data following the Cambridge Analytica scandal. Facebook is trying to quell those concerns with this MRI initiative, reassuring the public that "no Facebook data of any kind will be used in the project." 

There was also uproar earlier this year after CNBC reported that Facebook had been talking to hospitals about sharing data on patients, and the company subsequently said the project had been put on hold.

Join the conversation about this story »

NOW WATCH: This machine perfectly pours concrete

• Artificial intelligence poses both promise and perils for societies and their citizens.
• Kay Firth-Butterfield, who heads the AI program at the World Economic Forum, is working with governments, companies, and non-profits to help them understand the issues surrounding the disruptive technology, and work through ways to maximize its benefits and minimize harms.
• The most effective strategy for regulating AI will likely be through government and industry standard setting, she said.

If you ask Kay Firth-Butterfield about the promise and potential perils of artificial intelligence, she might start talking about toys, dolls, and action figures. 

Not the toys of today, necessarily, but those of the future, that will be empowered with AI. Such toys have the promise to be the quintessential educational tools, interacting with kids daily, gaining intimate knowledge about how they think and communicate, and using that information to help them learn.

"Personalized education using AI for kids is going to be a huge game changer," said Firth-Butterfield, who heads the artificial intelligence and machine learning program at the World Economic Forum, said in a conversation last week with Business Insider.

On the other hand, such toys raise a host of issues that policymakers are only starting to get a handle on. The privacy implications alone of potentially having a toy — or a succession of them — collect a child's every utterance from the time they can talk until adulthood are tremendous, she said.

"That is an issue that we really have to solve," she said.

And toys are just one of many areas where society will have to wrestle with both the potential and perils of AI.

The technology promises improvements to everything from industrial processes to agriculture to transportation, Firth-Butterfield said. But it also could lead to a raft of challenges and dangers, including massive job losses in a relatively short period of time, the illegitimate denial of goods or services thanks to flawed or biased algorithms, and citizens' loss of control of what was previously personal data, she said.

Firth-Butterfield works with governments and companies to think through AI issues

In her job with the World Economic Forum, Firth-Butterfield works with representatives of governments, corporations, civil society groups, and academic institutions to work through some of those challenges. The projects she and her group lead are designed to come up with ways to govern AI that will allow countries and companies to reap its benefits while minimizing its harms.

"It's really important that we know that there are all these different tensions, because without addressing them, we are really left with, I suspect, a failing trust in the technology," she said. "What I certainly don't want to see are all the benefits of AI somehow being lost because we haven't put in the ethical underpinnings to help the public know that we're doing something safe."

It's going to enable us to feed more people.

Firth-Butterfield, who has served as a lawyer and judge in the United Kingdom, has been helping people and companies work through the legal and ethical implications of AI for years now, as a professor, corporate advisor, and consultant. She joined the World Economic Forum last fall.

For her, AI has "enormous" potential. In education, it promises to provide students personalized learning programs that are tailored to their individual needs, learning styles, and aptitude.

In industry and business, the technology could help companies significantly reduce the amount of energy they use, she said. Google, she noted, announced two years ago that its DeepMind machine-learning technology helped it reduce the amount of energy it uses to cool the servers in its data centers by 40%.

And in agriculture, AI could be used in tandem with Internet of Things devices to make farmers and agribusinesses more productive and efficient, she said. AI could take the data collected by sensors in fields to help farmers determine how much fertilizers, pesticides, or even just water needed by their crops. 

"It's going to enable us to feed more people," she said.

AI has plenty of potential pitfalls

But she's equally concerned with the possible pitfalls of the technology. Algorithms that are flawed in design or in the data they rely on could lead to negative consequences for particular groups of people.

There's a long history of US lenders denying home loans to black people because of the color of their skin, for example. Software designed to automate loans approvals could end up perpetuating that prejudice if that bias is baked into the underlying algorithms, Firth-Butterfield warned. The same is true for discrimination in employment.

Harmful biases have already made themselves evident in artificial-intelligence software and tools. Two years ago, for example, the image recognition software built into Google Photos infamously labeled African-Americans as gorillas. Google also scuttled a video conferencing service intended for employees after the service's face-recognition software failed to detect the faces of people of color, Business Insider reported recently.

We don't have the luxury of a long time to actually even out the effects on job loss with this revolution, because it's happening so quickly.

"It's really important" that we make sure that we're "not encoding own prejudices and taking them forward with us, because if do that, we will actually stultify the development of the world," she said.

Biased algorithms aren't the only thing she's worried about. AI poses a big threat to employment worldwide.

The world had decades to adapt to the upheavals of the second industrial revolution, the one that led to mass production of everything from steel and automobiles, Firth-Butterfield said. But artificial intelligence is developing and likely will be adopted much more rapidly — and the impact on the job market will likely be felt in short order too, she said.

"We don't have the luxury of a long time to actually even out the effects on job loss with this revolution, because it's happening so quickly," she said.

And then there are the ways that AI could erode privacy and potentially harm kids.

Firth-Butterfield favors standards, not regulation

The best way to maximize the benefits of AI while minimizing the benefits is to have multiple stakeholders — governments, corporations, non-profit groups, and more — work through the issues and come up with ways to govern technology, Firth-Butterfield said. That doesn't have to be through laws and formal government regulations, she said. In fact, the better way to regulate AI will be to do it through government and industry standards, she argued.

By setting standards that attempt to minimize harms and take ethics into account, governments in particular can significantly influence the development of artificial intelligence, thanks in part to their huge purchasing power, she said. And setting standards tends to be a lot quicker and more flexible than crafting formal regulations or laws, so it can better respond to changing developments, she said.

AI's running fast, and we need to run as fast with governance mechanisms

"AI's running fast, and we need to run as fast with governance mechanisms," she said.

Those standards will need to focus on minimizing bias and protecting privacy, she said. They'll also need to make clear who or what entities are legally accountable for any harms that take place. And they'll need to ensuring transparency, so citizens and consumers understand how the AI algorithms work and what they're doing.

To be sure, there will be cases where governments will need to put in place formal regulations, Firth-Butterfield acknowledged. Those will likely be when they need to protect the most vulnerable people in society, including kids, the disabled, and the elderly, she said.

Already some countries are ahead of the game in thinking through AI governance issues, Firth-Butterfield said. Among them: Brazil, China, India, and the United Kingdom.

"There are a number of countries that are already stepping up to the plate," she said.

SEE ALSO: The founder of a beloved productivity app thinks the startup model is broken — here's how he's trying to keep the tech industry from 'making the same 10,000 mistakes over and over again'

SEE ALSO: A new study shows that tech CEOs are optimistic about the future, even if they still don't understand millennials

SEE ALSO: AI is great at recognizing nipples, Mark Zuckerberg says

Join the conversation about this story »

NOW WATCH: NYU professor says Facebook should pay taxes for making us less productive

• German scientists have developed software that can detect character traits through eyetracking, according to a new study.
• Using over 200 physical behaviours, such as the frequency with which subjects blinked, the researchers established connections between eye movements and personality traits.
• It's been suggested that the technology could eventually be developed to assist those with autism to gauge their peers' emotions and responses.

Anyone who's seen "2001: A Space Odyseey" will already be familiar with that nagging fear that Artificial Intelligence will one day become so intelligent, it will simply turn on us — but computer scientists from Saarbrücken and Stuttgart don't seem at all phased at the prospect.

According to a new study, they've developed software that can recognise personality traits through eye-tracking.

"There are already some studies in the field of emotion recognition through facial expression analysis. But this is the first time we've managed to establish personal traits," said Andreas Bulling in an interview with Business Insider.

Head of the research group "Perceptual User Interfaces" at the Max Planck Institute for Computer Science at Saarland University in Saarbrücken, he joined forces with scientists from Stuttgart and psychologists from the University of South Australia to develop software that can recognise character traits through eye-tracking.

The study found personal traits are more stable than emotions

As part of their research, the scientists had 50 students, all of whom were equipped with eye-trackers recording their eye movements, walk across their campus for around 10 minutes and buy something from one of the campus shops.

Subjects were asked to complete questionnaires commonly used to evaluate people's personality traits.

Using over 200 markers such as the frequency with which subjects blinked, how long they focused on something, and to what extent their pupils dilated, the researchers managed to establish which traits were associated with which eye movements.

Using the data obtained from the study, the scientists developed "decision trees" for various personality traits, after which the software was able to recognise those characteristics specified. Through eye tracking, the software can "see" whether someone is conscientious, sociable, tolerant, and even to what extent they might be emotionally unstable. It can also pick up on how inquisitive someone is.

However, what makes the AI even more special is that it doesn't require updates; rather, it's a "machine-learning" piece of software.

"It's novel in itself that this kind of software works at all," said Bulling. Emotions can vary from one situation to the next so they're often changing regularly and quickly; character traits, however, are very stable.

"Initially, it wasn't clear that it was possible to assess character traits from eye movements," said the expert.

The next step is to improve the software's performance: The experts want to focus on analysing body language too. The experts have expressed that, while it isn't their own vision to incorporate the software into hardware, there is great scope for the application of this software in the robotics field.

The software could even assist those with autism

"Like anything, what we've made can be used for both good and bad," said Bulling, but he hopes the software will lead to better and more authentic interaction between man and machine.

Just as an example, the eye-tracking software could eventually be installed in cars, which would then be able to recognise whether drivers are more willing to take risks or not.

If the software is sufficiently developed, it could even be used to assist people with autism: With eye-tracking glasses, those struggling to gauge their peers' reactions and feelings could better understand others' behaviour.

However, whether the software is used positively or negatively, Bulling said is not his responsibility: "That's ultimately for society and the politicians to decide."

Join the conversation about this story »

NOW WATCH: How a black cop infiltrated the KKK — the true story behind Spike Lee's 'BlacKkKlansman'

The Google Home is in millions of households. 

Earlier this year, Google revealed that it had sold more than one Home smart speaker device every second since October. According to our own calculations, that means Google sold at least 6.8 million Home devices during the last holiday season — and they seem to have been selling at a steady clip since. 

Google sells three Home devices right now: The $50 Google Home Mini, the $130 Google Home, and the $400 Google Home Max, all of which have Google's artificially-intelligent Assistant built in. 

So now that the Google Home has spread, it helps to know what you can actually do with them. Some features are obvious — like asking for the weather — but others aren't so obvious.

Here are 17 of the coolest things you can do with your Google Home. 

SEE ALSO: Your Amazon Echo can now send text messages for you — here's how to do it

1. Play white noise while you fall asleep.

I prefer rain sounds to standard white noise, so I usually say "Hey Google, play the sound of rain." The device obliges with a steady downpour. 

The sound usually lasts until I fall asleep, but if you want to be sure it turns off at a certain point, you can also set a sleep timer. 

2. Broadcast something to every Google Home device in your house at once.

If you have more than one Google Home device in your home, you can broadcast to every single one simultaneously, sort of like an intercom. 

If you say, "OK Google, broadcast that it's dinner time," each device will ring a dinner bell. You can also say, "OK Google, broadcast that it's time for school."

The broadcast feature even works when you're not in the house. Saying, "OK Google, broadcast I'm on my way home" will trigger the devices inside your house.

3. Control your smart home.

At this point, Google Assistant can control more than 1,500 smart-home products from more than 100 brands like LG, Whirlpool, GE, and Nest. You can ask your Google Home to dim the lights, change the temperature, turn on a kettle or microwave, or even start your Roomba. 

You can check out a full list of compatible products here. 

• HAL is a robot five-year-old boy.
• It can cry, call out for its mother, bleed, and get angry.
• Medical students can also practise all sorts of medical procedures on HAL.
• The idea is to add some emotional turmoil to medical training, helping the students to learn better.
• But HAL's developers didn't make it too realistic, in case it became too traumatic.

Medical training exercises are getting more and more realistic. Recently, companies have developed robots that medical students can practise on. The idea is that these pretend people can lead us a little way into the uncanny valley, so we have to deal with the emotional response as well as the methodology behind a procedure.

One of the latest medical robots is called HAL. It takes the form of a five-year-old boy which can respond to certain questions, follow a finger with its eyes, bleed, and convulse. It even has a pulse.

HAL was built by Gaumard Scientific, a company that produced the first synthetic human skeleton for medical schools. The company's technology has come a long way since then, having developed a synthetic boy who can simulate many medical problems, cry tears, and shout for its mother.

Using HAL is supposed to help students retain their knowledge better, because it is as close to treating a real person without actually using a human volunteer.

HAL's other functions include going into cardiac arrest, anaphylactic shock, and the ability to have its blood sugar, blood oxygen level, and carbon dioxide levels measured. Also, its pupils dilate when a light is shined into its eyes.

In a promotional video, a doctor asks HAL about how much its head hurts, and it responds "an eight."

To prepare for the really bad injuries and problems, HAL can be hooked up to real hospital machines and shocked with a defibrillator. When it's awake it can be set to several different emotional states, including lethargic, angry, amazed, quizzical, and anxious.

The idea is to make HAL just realistic enough to help students with their studies, but not so realistic that it's too traumatic to deal with when they have to slit its throat to insert a trachael tube.

HAL is one of a few medical robots currently in use. On the Gaumard website there is also a premature baby simulator, and a scarily realistic robot that gives birth.

These pretend people are very different from the lifeless dummies medical professionals have used for decades.

"I've seen several nurses be like, 'Whoa it moves!'" Marc Berg, the medical director at the Revive Initiative for Resuscitation Excellence at Stanford, told WIRED in a chilling article. "I think that's kind of similar to the idea that if you've driven a car for 20 years and then you got a brand new car, you're kind of amazed initially."

Watch the video explaining all of HAL's functions here:

SEE ALSO: Scientists have created a murder-obsessed 'psychopath' AI called Norman — and it learned everything it knows from Reddit

Join the conversation about this story »

NOW WATCH: Is marrying your cousin actually dangerous?

• Morgan Stanley says we've entered the type of computing cycle that comes around only every 10 years, with a specific focus on data technology.
• The firm has identified 26 companies it thinks possess a technological edge, whose stocks are thereby positioned to perform well in the future.

One way to achieve success as an investor is to identify trends before they fully develop.

You load up on shares of companies at the forefront to some sort of paradigm-shifting innovation, then watch in glee as you make money hand over fist.

Morgan Stanley realizes this and says it's actually a perfect time for traders seeking such opportunities.

At the center of the firm's latest analysis are the computing cycles it says emerge once every 10 years. It argues that this particular iteration has just gotten underway, with a focus on data technologies like artificial intelligence, Internet of Things, and automation.

"Following nearly two decades of underinvestment in technology, we see enterprises reinvigorating IT spend to enable productivity and believe data technologies support a secular tailwind to IT budget growth," Mike Wilson, Morgan Stanley's chief US equity strategist, wrote in a client note.

"Additionally, we see a clear mindset shift at the executive level from viewing technology as supporting the business to technology becoming the business," he added.

To that end, Morgan Stanley has identified 26 companies it says are best positioned to hold a technological edge over the coming years. By that rationale, they should make great long-term stock plays.

As the firm puts it, these companies are "at the cutting edge of a data era-driven productivity boom."

26. Extra Space Storage

Ticker: EXR

Industry: Real estate

Market cap: $13 billion

Source: Morgan Stanley

25. Spotify

Ticker: SPOT

Industry: Information technology

Market cap: $33 billion

Source: Morgan Stanley

24. MasterCard

Ticker: MA

Industry: Information technology

Market cap: $226 billion

Source: Morgan Stanley

• Kai-Fu Lee, a longtime technology executive who previously served as president of Google China, has helped US companies such as Microsoft and Google operate in China for years.
• As Google considers a return to China, he says the search business there is too mature and Google should give up on it.
• He says the situation with Waymo, however, is a different story.

If we're talking exclusively about Google's financial interests, then management's decision to pull out of China in 2010 was clearly a mistake, and Kai-Fu Lee, the former president of Google's operations in China, begrudgingly acknowledges that.

"Had they stayed and continued to operate legally, they would be in a much better position to launch products," Lee told reporters on Thursday. But the longtime tech executive and investor quickly noted that Google didn't leave China with an eye on its pocketbook.

The company was forced to choose between complying with a demand by China's government to censor search results, or stop operating in the country. Google's management chose to leave.

More recently, however, Google appears to have experienced a change of heart, and managers recently acknowledged they're considering a return to China. According to reports, Google is eager enough to return to China that it has already built a censored search engine.

The revelation that Google may bow to China's censorship demands has drawn criticism from human-rights groups, US politicians, and even some of the company's employees. Lee, who worked for Google China for four years before leaving in 2009, made his comments following a panel discussion he participated in at the Artificial Intelligence 2018 Conference in San Francisco.

'In search, it's just not meant to be'

Lee declined to say whether he thought the stand Google took nine years ago for free speech was the right one. He did, however, have no problem listing the many competitive reasons Google might not want to reenter China.

For starters, the search business there is packed with homegrown competitors, whom Lee earlier likened to entrepreneurial "gladiators." According to Lee, Google shouldn't even bother facing off against this lot because the odds are all with the home team.

"When you're in a market that is already 20 years old and mature like search, to go in now with a zero market share and build it up is such an uphill struggle," Lee said. "When I went into the search market it was maybe eight years old and Google had 9% market share, so we had something to work with and a brand name."

Lee said that before Google left China, he and his staff had carved out a 24% share of the market. He's doubtful Google could come close to that again.

The search giant, however, would stand a better chance of finding success if it moved into a segment that wasn't as mature, Lee said — say, for example, autonomous vehicles.

"In search, it's just not meant to be," Lee said. "If Google wants to do anything, they should enter an area with a new product in which there are no entrenched players, nor are there clear user expectations and biases. For example, if Waymo could land in China, Google would have such advantages. A two-year lead and also if there's proper deployment, it could be a runaway success."

SEE ALSO: A top Google researcher gave a spooky talk about how Silicon Valley companies could one day read and store our thoughts

SEE ALSO: Google's former China boss says the search company won't stand a chance against today's Chinese 'gladiator' entrepreneurs

Join the conversation about this story »

NOW WATCH: Everything Samsung just announced — the Galaxy Note 9, Fortnite, and more

• Former Google China president and legendary tech investor Kai-Fu Lee says China is outpacing the US in the development of its mobile applications and AI.
• He predicts that the country's tech will usurp the US in the fields of retail, education, and healthcare in upcoming years.

When Kai-Fu Lee, prominent Chinese tech investor and former president of Google China, considers some of the leading technology products in the United States, he's underwhelmed.

"Look at YouTube or Instagram or Snapchat," said Lee in an interview with Business Insider. "If China had one of those companies and it moved at the pace that those companies move, it would likely be dead in the water."

For decades, Chinese developers closely watched America's tech sector for signs of innovation. But now, with its fast-developing mobile market breaking new ground, the country's tech sector is swiftly gaining confidence.

According to Lee, China has already outpaced the US in terms of creating better mobile applications and mobile payment systems. In the future, he predicts that technologies in the fields of retail, education, and healthcare will all soon leapfrog ahead of their US counterparts.

Notably, said Lee, Chinese entrepreneurs aren't interested in being followers.

"Chinese entrepreneurs are at a stage where they no longer feel they have to copy anyone else," said Lee. "They have enough experience and knowledge of the market and know how to build a deep, long moat around their castle."

This so-called "castle" of China's burgeoning tech sector is among the fastest developing markets in the world, and Lee predicts that it's on track to outpace the US, particularly in the area of artificial intelligence. 

Among the 13 companies Lee has invested in that are worth $1 billion or more, five are artificial intelligence companies. Together, said Lee, these five companies make up a cumulative value of $23 billion. 

See more: 'The future will be won or lost on this technology. I'm very concerned': The co-founder of a $9 billion company warns that China is on track to dominate the US in AI

There's several factors at play in the country's flourishing tech economy.

For one, Chinese companies simply innovate differently than those within the US. 

Lee offered up an example of how video has evolved within the country.

"China pushed the market to change from longform videos to short-form videos," he said. "From there, it changed from videos being watched to videos that you're apart of, and then to real-time videos, to using videos as a form of social networking. There is always someone trying to iterate and do something more."

"In China you assume from the start that someone is coming after you"

Chinese entrepreneurs think differently, as well, said Lee. 

"The business practice in China is 'winner takes all,'" said Lee. "They're not constrained by cornering only a sliver of the market. It's not about having a niche. The entire eco-system is interdependent."

This ecosystem differs from the way US entrepreneurs innovate, said Lee. 

"The American system is a gentlemanly system that involves respect, innovation, and consensus," said Lee. "People work in teams. They brainstorm and go into stealth and then come up with something."

But in China, entrepreneurs are less preoccupied with cultivating a creative, communal process.

"China is a giant experimental ground where people test things out," said Lee. "If they don't work they keep iterating over and over again. They want to be first. In China, you assume from the start that someone is coming after you."

SEE ALSO: People are talking about this amazing photo of Jack Dorsey and Alex Jones as a funny, dystopian, yet iconic image

Join the conversation about this story »

NOW WATCH: Why TV speakers suck

• British startup Babylon Health says it will invest $100 million from its own balance sheet into hiring hundreds of scientists to push forward the use of artificial intelligence in health.
• Babylon already has several contracts with the NHS, the UK's national health service, and its AI medical service is embedded in Samsung phones in the US.
• The company will move to a new headquarters in London within the next 18 months to house its new staff.
• CEO Ali Parsa was outspoken in his rejection of criticism from doctors, who have pointed out that Babylon's claims are not peer-reviewed and that its "GP at Hand" app may be having a negative impact on GP surgeries.

British medical startup Babylon Health will invest $100 million in hiring more than 500 researchers, scientists, and engineers over the next year to develop the use of artificial intelligence in healthcare.

Chief executive Ali Parsa said this would more than double Babylon's current team to more than 1,000 staff, and that the company would eventually move into new headquarters in London over the next 18 months to house the new recruits. The firm is currently headquartered in Kensington, London, and is planning to take over the rest of the offices in its building. 

The artificial intelligence research will build on software that Babylon showed off in late June. The company claimed at the time that its AI could assess common conditions more accurately than doctors. The idea is to expand the AI from primary care, which refers to the stage when patients first approach their GPs, to helping to diagnose and manage chronic diseases, which is currently threatening to overwhelm the UK's NHS health service.

Babylon's announcement coincided with an endorsement from health secretary Matt Hancock, who has pledged to reform the NHS with technology.

But the company has come under considerable scrutiny and criticism from the medical community.

Specifically, doctors have criticised the "GP at Hand" app, which is powered by Babylon and offers patients video appointments through an app. The worry is that Babylon would receive NHS funding for looking after younger, fitter patients who would be more likely to use a digital service. That would leave physical GP surgeries with the burden and expense of looking after elderly patients with complex needs.

They have also criticised Babylon's June claims about its AI chatbot, namely that its claims had not been peer-reviewed. Babylon did publish an explanation about how its AI worked on its site, though it's listed under the "marketing" section.

In a call with Business Insider, chief executive Ali Parsa dismissed the criticism that Babylon was cherry-picking younger patients as "complete nonsense." 

He said not only were older patients increasingly using GP at Hand — though he didn't give figures — but added that surgeries are paid less for looking after younger, fitter patients. "The cherries are actually the elderly," he said.

Asked about whether Babylon would submit its AI research for peer review, he said that model of waiting 18 months for submissions to be accepted to an academic journal was outdated.

He added that patients themselves were satisfied with the technology.

"One person every few seconds is using our technology," he said, citing a net promoter score of 82, a metric usually used by marketing departments to measure customer satisfaction. "We published the paper, we showed the methodology... the only thing we didn't do was this situation  of waiting 18 months for a peer-reviewed paper."

Parsa said the company had to submit its paperwork to medical organisations such as the MHRA and the US Food and Drug Administration. "Everybody can see our methodology but this peer-review model ... which two giants of journals advocate massively, was created in the 19th century, when science moved at snail pace." (Scientific American suggests the birth of peer review was in 1731.)

Parsa's criticism is unlikely to win over the cynics. His view is that technology moves too quickly to wait for a lengthy evaluation process.

Asked if the startup was now profitable, Parsa said licensing its technology to partners such as Samsung and Chinese internet giant had been hugely lucrative, and said the firm was in and out of profitability. He said the firm was focused on growth and expanding its capabilities.

SEE ALSO: Investors backed an AI startup that puts a doctor on your smartphone with $60 million

Join the conversation about this story »

NOW WATCH: Apple might introduce three new iPhones this year — here’s what we know

• Nvidia received a $350 price target at Needham — the most bullish on Wall Street. 
• The chipmaker on Wednesday launched AI data center platform TensorRT, which could deliver the industry’s most advanced inference acceleration.
• Nvidia last month unveiled a new lineup of video cards, GeForce RTX series, which can perform real-time ray tracing thanks to its new Turing architecture. 
• Watch Nvidia trade in real time here.

Nvidia shares jumped more than 2% early Friday after Needham raised its price target to $350 a share — the most bullish on Wall Street — saying its dominance in artificial intelligence and has it well positioned to capitalize on the next wave of computing.

"We see striking parallels between Nvidia's dominance in Artificial Intelligence/ Machine learning and the 'Wintel' platform during the era of PC computing," Needham analyst Rajvindra Gill in a note sent out to clients on Friday.  Wintel comes from two words "Windows" and "Intel" and refers to all personal computers based on the Intel microprocessor and one of the Windows operating system from Microsoft.

Join the conversation about this story »

NOW WATCH: One bite from this tick could ruin red meat for the rest of your life

• Amazon lately has been bulking up the sports-related knowledge of its Alexa intelligent assistant.
• Alexa can now give the odds on upcoming games and answer questions about players' career statistics.
• In coming months, it will also be able to notify users when a game is about to start and offer updates on users fantasy football players.

In the past, if you wanted to know what was going on in the sports world, you might have tuned in ESPN or headed to your favorite sports website.

Now you might want to just ask Alexa, Amazon's intelligent assistant, best known as the voice of the Amazon Echo line of speakers. 

Just in time for the NFL season, Amazon has been stuffing Alexa full of sports knowledge. You can now ask it to give you the odds on the upcoming game between the Oakland Raiders and Denver Broncos, to tell you how many passing yards Joe Montana threw for in his career, and to give you the name of the Pittsburgh Steelers backup running back. In the near future, Alexa will be able to give fantasy football fans updates on their players, and alert users when their teams are about to take the field.

Sports-related questions have become some of the most popular ones to ask Alexa in recent years, Jason Semine, principal product manager for sports information on Amazon's Alexa team, said in an email.

"Sports moments are very important to our customers," Semine said. He continued: "So, this year we wanted to double-down in this area and see where else Alexa might be useful as it relates to sports."

In recent weeks, Amazon has added a slew of new sports-related features to Alexa. Among them:

• Answers to an assortment of trivia-related questions relating to sports history, records and statistics.
• Updates on the latest injuries and transactions involving individual players or teams.
• Predictions on who will win upcoming games, including the latest betting line.
• Recaps of completed games.
• Radio-like reports offering fans the latest information on their favorite teams.

Within the next several months, Alexa users will also be able to ask for updates on their fantasy players and have it notify them when particular games start.

Alexa could know even more about sports soon

Alexa relies on some 50 different sources, including Stats and Sportradar, to provide answers to sports-related inquiries, Semine said. But Amazon has done a lot of work in the sports area in-house, including helping Alexa better respond to questions that are phrased using regular language rather than with specific command words.

The Alexa team has also been working to ensure that the intelligent assistant is able to respond to questions about sports events as they happen and to understand the context of particular inquiries, he said. If an Alexa users asks for the score of the Stanford game, the team has been trying to fine-tune the system so that it knows whether the user is referring to football, men's basketball, or women's basketball when you ask about a game.

"The Alexa service is getting smarter every day," Semine said. He continued: "Our long-term goal is for Alexa to understand and be able to answer all questions, in all forms, from anywhere in the world."

Alexa isn't the only intelligent assistant that knows about sports. Apple's Siri has long been able to answer sports-related questions, including ones about team standings and player-related statistics. Earlier this year, the company added to it the ability to respond to questions about professional tennis and golf.

Meanwhile, Google Assistant can answer a selection of sports queries, including about teams' latest scores and standings. Apps, or skills, that connect with Assistant can allow the service to respond to an even broader range of inquiries, including ones related to your fantasy football players or the latest horse-racing results.

SEE ALSO: A gold mine is buried 'under the weeds' at Amazon — here's why it could take the company beyond the $1 trillion mark

SEE ALSO: Apple is overhauling Siri — and the voice assistant may finally catch up with Amazon's Alexa

Join the conversation about this story »

NOW WATCH: An environmental group is testing giant floating pipes to clean up oceans

• Sea mines remain a real and terrifying threat and are a central part of many countries' plans to disrupt maritime operations in the event of a conflict.
• The US Navy has been looking at remote and autonomous systems to help counter mines at sea for some time.
• The Marine Corps is also looking for AI and other systems to spot and neutralize mines near the shore in order to ensure the success of amphibious landings.

After nearly two decades of fighting in Afghanistan and Iraq, the Marine Corps is looking to reorient toward its specialty, amphibious operations, while preparing for the next fight against what is likely to a more capable foe.

Peer and near-peer adversaries are deploying increasingly sophisticated weaponry that the Corps believes will make amphibious landings a much more challenging proposition in the future. 

The Corps is looking for high-tech weapons to counter those looming threats, but it's also looking for a sophisticated system to counter a persistent, low-tech, but decidedly dangerous weapon — mines hidden close to shore.

According to a recent post on the US government's Federal Business Opportunities website, first spotted by Marine Corps Times, the Marine Corps Rapid Capability Office is looking to autonomous and artificial-intelligence technology to "increase Marines' ability to detect, analyze, and neutralize Explosive Ordnance (EO) in shallow water and the surf zone"— two areas where amphibious ships and landing craft would spend much of their time.

"Initial market research has determined multiple technically mature solutions exist that can assist Marines ability to achieve this capability," the notice says.

Potential systems envisioned by the Corps' request for information include autonomous or remotely operated vehicles, unmanned underwater vehicles, and unmanned aerial vehicles outfitted with sensors and other gear to detect and evaluate explosive devices.

"Some solutions may provide the ability to neutralize detected ordnance, which is desired but not required," the RFI states.

The Corps wants contractors to submit up to three prototypes from a single family or multiple families of systems.

Requirements outlined in the RFI for contractor-submitted systems include being able to detect and identify explosive devices in waters ranging the surf zone, where depths are less than 10 feet, to very shallow waters, which range from 10 feet to 40 feet in depth.

The proposed system must also be able to navigate and avoid obstacles in the littoral zone, which includes shorelines out to coastal waters of 200 feet in depth or more.

The system submitted to the Corps must also be able to use geolocation information to "mark" explosive devices to within a meter in environments where communications and GPS are contested or denied.

The Corps is also looking for systems that are man-portable and can be launched and recovered by one- or two-man teams in a small boat, like the Combat Rubber Raiding Craft.

While mines have grown more sophisticated in recent decades, even rudimentary ones are still a potent threat.

An Iranian sea mine that almost sunk the US Navy frigate Samuel B. Roberts in 1988 was a World War I-era device. Since the end of World War II sea mines have destroyed or damaged more US Navy ships than any other weapon.

Mines have become a cornerstone of anti-access/aerial-denial strategies adopted by countries like Iran and China, which have plans to deploy them in important maritime areas like the Strait of Hormuz or the South China Sea.

The Navy has dedicated mine-countermeasures systems, including specially designed and equipped Avenger-class ships that are deployed around the world and rapidly deployable MH-53H Sea Dragon helicopters that often accompany Avenger-class ships.

Those systems are aging, however, and the Navy has been working on a slew of remotely operated and unmanned mine-countermeasures systems that would be deployed aboard the service's littoral combat ships, with the goal of "taking the man out of the minefield."

While there has been recent progress with LCS-based anti-mine systems, the LCS program and those mine countermeasures have encountered delays, malfunctions, and cost overruns that have hindered the program and its implementation.

The Corps has also made progress with countering mines that Marines would encounter on shore.

In December 2017, Marines conducted the first amphibious landing with a modified full-width mine plow prototype, which was attached to an assault breaching vehicle and sent ashore on during an exercise on the West Coast.

The regular full-width plow was too big to fit aboard the Navy's landing craft utility boats. The modified version is easier to transport and safer to use, a Marine Corps Systems Command official said earlier this year, and it gave commanders more flexibility with their ABVs.

Once ashore, the plow supplements the ABV's other mine-countermeasure systems, helping clear a path for Marines to advance off the beach.

"This plow prototype makes the ABV transportable and gives the commander options to accomplish his tasks on the battlefield," Alvin Barrons, an assault breaching vehicle engineer, said in a release at the time. "The capability makes the force more lethal because it helps keep other combat vehicles intact and saves the lives of Marines."

SEE ALSO: The US Navy faces 'a huge liability' in countering one of Iran's favorite and most dangerous weapons

Join the conversation about this story »

NOW WATCH: Step aboard the USS Kearsarge, the US Navy workhorse that takes Marines to war

To hear the experts tell it, if your business is connected to healthcare, energy, financial services manufacturing or call centers, then there's a chance you'll soon be toast.

A group of companies, including Google, Amazon, Apple, Microsoft, Oracle, IBM, and Intel are investing big money to develop artificial intelligence and other game-changing technology. These new AI superpowers are expected to pay your industry a visit in the not-so-distant future, leaving key parts of your product obsolete and the economics of your business in tatters.

The good news is that some people have faced the dramatic changes brought on by a technological tsunami  and seen their sectors not only survive, but eventually flourish again.

Cary Sherman, CEO and Chairman of the Recording Industry Association of America, the lobbying group for the three top music labels, is one of those survivors. 

The advent of digital music files and online sharing that started in the late 1990s plunged the music industry into its darkest period — in 2009, at the low point, US annual music sales fell to $6.3 billion, less than half of the $14.6 billion posted for 1999.

Now, nearly a decade later, annual music sales are growing again and consumers are paying for streaming services that provide access to an unprecedented trove of music. 

It wasn't easy. Sherman and the record labels endured numerous false starts, surprises and setbacks. They erred. The number of labels fell from five to three. They laid off hundreds of workers and alienated fans. They were mocked and vilified by the tech press

What follows is the story of how recorded music clawed its way back. Sherman's account is a playbook of tips for other industries that may soon find themselves in the fight of their lives:

SEE ALSO: EU lawmakers vote for new online copyright rules

1. Remember, it may take time to grasp the nature of the threat

Napster, the file-sharing service cofounded by Shawn Fanning, launched in June of 1999.

Sherman and his staff immediately recognized it as a threat but didn't appreciate how big of a threat it was.

"I think everybody underestimated the speed at which it would grow and the size that it would grow to," Sherman tells Business Insider. 

Before a court ordered Napster shut down in 2001, the service had enabled maybe as many as  8o million people around the world to share countless numbers of songs, free of charge. 

If a 19-year-old college kid with minimum resources could create a worldwide phenomenon and flip the music industry on its head, what might be accomplished by seasoned technologists who were backed by big bucks? 

2. Accept that the competitive threat may evolve, morph and quickly adapt

Napster was a centralized peer-to-peer service. A central server indexed the users and their libraries of MP3 song files so others could access them. The music industry had seen earlier, clunkier, and far less popular versions, called File Transfer Protocol technology. 

"We started with FTP sites," Sherman said. "Napster was the next form of piracy in 1999 and then a couple of years later came a decentralized form of piracy. Then, came cyber lockers and so on… each generation of formed piracy had a completely different scale beyond anything we had seen before. And so we saw very quickly that things were spinning out of control." 

3. Expect your entry barriers and moats to be breached

The year before Fanning invented Napster, the big recording companies thought they had piracy-proofed their business thanks in large part to the Digital Millennium Copyright Act, which outlawed the cracking of anti-piracy protections on CDs and the distribution of pirated songs by Internet services. 

But the new breed of services let users store music files on their computers, rather than on centralized servers owned by internet providers. That meant that internet service providers weren't liable for the pirated music and it left the record labels with no effective way to stop the problem. 

The DMCA, the record industry's main bastion of defense against piracy, was "obsolete within eight months," says Sherman.

• Alibaba aims to launch its first self-developed AI inference chip in the second half of 2019 that could be used for autonomous driving, smart cities and logistics.
• The company will set up a dedicated chip subsidary as well. 
• Alibaba's aggressive drive to develop its own semiconductors comes as China's government looks to raise the quality of home-made chips to help propel high-tech domestic industries from cutting-edge transport to AI healthcare systems.

HANGZHOU, China (Reuters) - Alibaba Group Holding Ltd will set up a dedicated chip subsidiary and aims to launch its first self-developed AI inference chip in the second half of 2019 that could be used for autonomous driving, smart cities and logistics.

The Chinese firm said at an event in Hangzhou on Wednesday that the new subsidiary would make customized AI chips and embedded processors to support the firm's push into fast-growing cloud and internet of things (IoT) businesses.

Alibaba's aggressive drive to develop its own semiconductors comes as China's government looks to raise the quality of home-made chips to help propel high-tech domestic industries from cutting-edge transport to AI healthcare systems.

In April, Alibaba bought a Chinese microchip maker Hangzhou C-SKY Microsystems to help bolster its cloud-based "internet of things" (IoT) business.

Jack Ma, Alibaba co-founder and chairman, said then that China needed to control its "core technology" like chips to avoid over-reliance on U.S. imports, something which has been put in the spotlight by whipsawing trade tensions.

(Reporting by Cate Cadell; Writing by Adam Jourdan; Editing by Darren Schuettler)

Join the conversation about this story »

• Gmail's "Smart Reply" feature suggests quick responses at the bottom of emails. 
•  "I love you" was often a suggested reply until Google changed it in testing. Another common early suggestion was "Sent from my iPhone."
• The current version of Smart Reply doesn't often suggest these responses. 

A few days ago, I received a short, effective email in my inbox: "Sounds good!" 

I had to pause. Although that was the response I wanted — I was arranging a meeting — I wondered: Did he really send that, or did he simply hit Google's automated response suggestions at the bottom?

In recent weeks, Gmail's "Smart Reply" feature has been released to an ever-rising amount of the web email service's over 1.4 billion users, meaning that soon a large percentage of the world's population will get access to Google's cheery, direct suggestions when it becomes a default feature next month. 

The actual language of the suggestions is created with machine learning, a family of technologies that is sometimes called artificial intelligence in the non-technical world.

Google's software crawls billions of emails on Gmail, and then uses software to suggest phrases drawn from that database that you might want to use to confirm a coffee at 4 PM at Blue Bottle. 10% of responses are generated through the Smart Reply feature, according to The Wall Street Journal. 

But it hasn't always worked well enough — a director of product management at Google, Ajit Varma, shared some funny early bugs with the Journal. 

From the story:

Google said an early prototype of the feature had “a propensity to respond with ‘I love you’ to seemingly anything,” forcing it to tweak the algorithm. “You don’t want to respond that to your boss,” Mr. Varma said.

Imagine trying to organize a conference call with a client and accidentally sending "I love you." 

Another funny suggestion the software kept repeating is a nod to Google's neighbors in Silicon Valley, Apple. Apparently, the software thought "Sent from my iPhone" was a good way to respond to emails.

That's the default signature on the default email app on iPhones, so it shows up all the time in Google's data. And Google's machine learning software, as smart as it was, isn't actually a human, so it didn't realize that was a clever bit of branding, and not any kind of useful information. 

Same with "I love you"— while Smart Reply is impressive, it doesn't understand social context yet. 

But the software does learn: Google says that as you use it more, the style and tone will become more personalized to you. 

Sounds good! 

SEE ALSO: China and Sweden enter bizarre feud after tourists throw themselves on ground claiming 'brutal' murder attempt

Join the conversation about this story »

NOW WATCH: Watch Apple unveil a new, bigger watch

• Facebook is on a mission to help people who speak different languages understand one another.
• It's using M Suggestions, its virtual assistant in Messenger, to translate real-time conversations, and has just added support for French.
• Business Insider spoke to Facebook employees working on the project about the tech behind it, and its potential to radically affect online communication.

In 2015, virtual assistants were the Next Big Thing.

Every major tech company had one, from Apple's Siri to Amazon's Alexa and Google Now, which would later become the more fully fledged Google Assistant. Facebook entered the space in August that year with the announcement of M— a chatbot that lived inside its Messenger app, and which users could ask for just about anything, from hiking recommendations to help buying flowers. 

But while the likes of Alexa and Google Assistant have exploded over the last few years, finding their way into everything from fridges to cars, M — in its original incarnation — was never widely rolled out to Facebook's users. 

Instead, it morphed into M Suggestions, an AI enhancement that hovers inside users' Messenger chats with their friends and offers contextual suggestions based on their conversations — from making payments to initiating video calls. These recommendations have historically been relatively incremental — send a funny GIF! Attach a sticker! — but Facebook is now leaning hard into translations in Messenger, and using M Suggestions as the engine to do so.

It's by far the most significant application of the technology behind M Suggestions to date — one that has the potential to radically reshape how hundreds of millions of people communicate in real-time on Facebook. Business Insider spoke to some of the team at Facebook working on the project, to learn more about the tech underpinning the ambitious project and their vision for its future.

"At Facebook we have a lot of different cultures and a lot of diversity in our team ... we are all speaking different languages, and we know how frustrating it can be not to be able to communicate ... in your own languages," said Laurent Landowski, a Facebook product manager. "So being able to also really provide M Suggestions for translation to the world is something that we're super proud of."

M is an AI assistant that lives inside other conversations

Every time you send a message to a friend via Messenger — assuming it's not an encrypted "Secret Message"— it's being scanned by Facebook.

That doesn't mean an actual Facebook employee is reading it, of course. Instead, Facebook's automated systems are parsing the message, trying to understand the intent of the message. This effort is partly underpinned by the tech Facebook acquired when it bought natural language startup Wit.AI back in 2015; cofounder Landowski is now the product manager of M Suggestions at Facebook.  

If one of Facebook's AI neural net models identifies the message between you and your friend as something it can add context to, M Suggestions will automatically spring into action. If you mention a song, it might prompt you to play it on Spotify; if you're discussing multiple potential activities in a group, it might suggest creating a poll.

And it learns over time what different people utilize it for, and caters its responses accordingly; the version of M that a GIF-addict sees will be very different to what appears for someone who is more restrained in their messaging.

When it comes to quick replies — suggested responses M offers users in conversations to save them time — it even uses users' conversation histories as a training data to teach the AI how they speak, making the responses ("yes" versus "yeah" or "yep" or "yah") sound authentically like their voice.

Its unique positioning — inside users' existing conversations with other people, rather than dedicated human-to-AI chat windows — means it risks being invasive or jarring. As such, Facebook has moved slowly adding suggestions to M, Landowski said. "We have been working on trying to improve and really focus on the delight and relevance as opposed to the number of suggestions we could be suggesting."

He added: "It is super easy to lose user trust."

Translations aren't easy — but the pay-off is huge

M Suggestions has thus far offered fun enrichments to conversations, but it's hardly transformative. Where that changes is translations.

Facebook has provided language translations on its core social network for years, first via traditional phrase-based translation techniques before migrating to a more advanced AI-powered neural net translation system in August 2017. 

However, users who wanted to be able to talk across language barriers in real time were out of luck until earlier this year, when Facebook launched the first Spanish-English translations in Messenger, underpinned by M Translations. The company followed it up with the announcement this week that it was adding French.

While Facebook positions Messenger at least in part as a way to stay in contact with the people users are closest to, translations opens it up to assisting people who may never have interacted before — like in Marketplace, Facebook's peer-to-peer sales platform. 

"You see more and more ways where translations can be applied, not only in your personal messaging but also like in the Marketplace, buyer-and-seller-type of messaging, that can unlock a lot of further opportunities for use," said Landowski, a native French speaker from Paris.

But translation isn't easy. Languages are always changing and shifting, evolving as slang becomes common parlance, and the problem is especially acute on an informal, real-time platform like Messenger. Facebook's language team has an "active taskforce working to adapt its models to the type of data that Messenger provides, said Necip Fazil Ayan, head of Facebook's language and translation technologies team.

"This is one of the hardest problems we have to deal with while working on translations at Facebook, and it's not a solved problem, making our systems more robust to informal language including slang," he said.

"It's a dynamic language right, and people keep inventing stuff ... my best example instead of just saying 'happy birthday,' [they] start introducing lots of P's or Y's or I's all over the place."

Facebook's unprecedented digital archives of billions of users' public and private conversations means it has a vast dataset with which to train its AI — but some languages have more material available than others. "One of the biggest challenges we're dealing with, both in terms of language understanding and translation perspective, is the set of what we call 'low resource languages.' As you can imagine, all the machine learning models require a lot of data to become accurate ... and we don't have that luxury for a lot of the languages we are dealing with," Ayan said.

And then there's the issue of bias. AI systems are only as good as the data they are trained on, and when there are human biases built into the data, it can creep into the results. "Our data is biased ... we are actively working on this at Facebook ... in machine learning this is a very hot area and it's a very difficult area, and it's going to take a while cleaning up the data from that type of bias or learning where the bias is." 

But for all the challenges, automated real-time translations offer Facebook a way to have a have a fairly profound effect on the way people around the world communicate. "I really don't want language to become a barrier when people are expressing their opinions or when people are trying to reach other people to get different perspectives," said Ayan, a native Turkish speaker who grew up in the country.

"So that's the dream, we are going to break down the language barriers and that's my personal mission here."

'It's hard from an AI perspective to be able to create a fully automated assistant that can do everything'

Facebook now describes the original version of M as an experiment, one that provided valuable insight into the kind of things users utilise chatbots for but was never intended as a competitor to Siri et al, and never scaled beyond 2,000-odd users in the Bay Area. (In contrast, more than 100 million people interact with M Suggestions a month as of November 2017, a spokesperson said.)

"It's hard from an AI perspective to be able to create a fully automated assistant that can do everything,"Landowski said, "We basically realized that people, especially in Messenger, they really want to focus on the communication [assistance that M offered] ... it was really about trying to be where they are, which is their actual conversations. Instead of talking to an assistant directly it was all about focusing on the communications."

But there have been rumours circulating for months about Facebook building a smart speaker in the vein of the Amazon Echo of Google Home, with a voice-controlled AI assistant built in.

It was a no-show at Facebook's annual F8 conference in May 2018, but speculation was bolstered after a reverse engineer found references hidden in Facebook's code for an "Aloha" voice feature.

"We're exploring everything," Landowski said when asked about whether Facebook was currently looking at speech recognition.

"Speech is also an interesting way to actually interact with an assistant. We cannot comment more on exactly what we do and what we test, but definitely we are working and trying to improve this."

Contact this reporter via Signal or WhatsApp at +1 (650) 636-6268 using a non-work phone, email at rprice@businessinsider.com, WeChat at robaeprice, or Twitter DM at @robaeprice. (PR pitches by email only, please.) You can also contact Business Insider securely via SecureDrop.

SEE ALSO: 2 brothers who worked at Facebook are on a mission to help engineers avoid terrible startups and make a boatload of money

Join the conversation about this story »

NOW WATCH: Apple's entire iPhone XS event in 8 minutes

• Artificial intelligence (AI) is developing rapidly and starting to be adopted widely.
• The technology has the potential to transform society.
• But it could lead to lots of negative outcomes, such as massive unemployment, and could be put to plenty of deleterious uses, such as large-scale violations of privacy.
• The best way to avoid those harms would be to focus on creating products that use the technology in socially beneficial ways, said Phil Libin, best known as the founder of Evernote, whose new startup, All Turtles, incubates AI projects.

When it comes to how artificial-intelligence technology might affect society, there are a host of things to worry about, including the massive loss of jobs and killer robots.

But the best way to avoid such negative outcomes may be to ignore them, more or less.

That's the advice of Phil Libin, CEO of All Turtles, a startup that focuses on turning AI-related ideas into commercial products and companies. In a recent conversation with Business Insider, Libin likened the situation to some advice he received when he was learning to ride a motorcycle.

His instructor taught him that if an accident happened in front of him while he was riding on the highway, such as a semi truck flipping over, the worst thing to do would be to stare at the truck. Instead, his instructor said, he should focus on the point he needed to get to to avoid colliding with the truck.

"If you look at what you're trying to avoid, then you're going to run into it," said Libin, who previously founded Evernote. "You've got to look at where you want to be."

The tech industry would do well to follow that admonition when it comes to developing artificial intelligence, he said.

Years in the making, AI is starting to progress rapidly. It's being used by consumers in the form of intelligent assistants such as Amazon's Alexa to answer trivia questions and make purchases. And it's being used by corporations to help them make business decisions.

AI has the potential for good — and evil

Many observers think the technology could transform society in profound ways, and not necessarily for the better.

Indeed, there are some potentially dangerous and dystopian outcomes and uses of AI. It's already starting to be used in China as part of a mass surveillance scheme. It could be used to track people basically from their birth on, collecting intimate insights into their every thought and desire. It could be used to perpetuate or worsen discrimination against particular people or groups. And it could be used to power terrifying new weapons.

Technologists and policy makers ought not ignore such potential uses of the technology, Libin said. They should be aware of them. But the best way to avoid them would be to concentrate on developing ways to use AI in socially beneficial ways, he said.

"There really is a flipped-over truck, and there's all sorts of bad things that can happen. And we definitely need to work towards not hitting it," Libin said. "But the best way to do that is to [say] … this is where we want to go. Here's a vision of certain products that are like obviously good, and virtuous, and the world needs them, and they solve real problems, and let's make those products."

Indeed, that's what he sees as a big part of All Turtles' mission. One of the first projects the company helped incubate is a chatbot called Spot that is designed to make it easier for employees to document and report incidents of sexual harassment and discrimination. Another is Disco, a plug-in for collaboration software Slack that helps employers give timely positive feedback to workers.

The projects All Turtles works on "is all stuff that we should be able to, right from beginning, right by design, feel good about," he said.

SEE ALSO: This former judge is heading the World Economic Forum's approach to AI — here's why she thinks regulation is unlikely, and what should be done about AI instead

SEE ALSO: The founder of a beloved productivity app thinks Hollywood has a better blueprint for innovation than Silicon Valley —and he's taking his cues from Netflix to fix it

SEE ALSO: The founder of a beloved productivity app thinks the startup model is broken — here's how he's trying to keep the tech industry from 'making the same 10,000 mistakes over and over again'

Join the conversation about this story »

NOW WATCH: British Airways has a $13 million flight simulator that taught us how to take off, fly, and land an airplane