Speaking at the Carnegie Mellon University – K&L Gates Conference on Ethics and AI in Pittsburgh, Eric Horvitz made the revelation. He says the group at Microsoft looking into possible misuse on a case-by-case basis is the Aether Committee (“Aether” stands for AI and Ethics in Engineering and Research.)

“Significant sales have been cut off,” Horvitz said. “And in other sales, various specific limitations were written down in terms of usage, including ‘may not use data-driven pattern recognition for use in face recognition or predictions of this type.’”

Horvitz, of course, did not reveal the specific companies of which Microsoft decided not to formulate a deal with. However, it’s pleasing to hear the company putting ethics above money when it comes to artificial intelligence. Abuses will be widely covered and hamper its potential.

Amidst the fallout of the Cambridge Analytica and Facebook scandal, and the use of this stolen data to target voters during the 2016 U.S. presidential campaign, people are naturally more wary of anything which involves mass data analysis.

Manipulating votes is one of the key concerns Horvitz raises for the abuse of AI, along with human rights violations, increasing the risk of physical harm, or preventing access to critical services and resources.

In reverse, we’ve already seen how AI itself can be manipulated — from Microsoft itself, no less. The company’s now infamous ‘Tay’ chatbot was taught by people online to spew racist comments. “It’s a great example of things going awry,” Horvitz acknowledged.

Rather than replace humans, Horvitz wants AI to be complementarity and not a replacement — often as more of a backstop for human decisions. However, it could still be used when invoked for tasks where a human would not be as effective.

For example, Horvitz highlights a program from Microsoft AI that helps caregivers to identify patients most at risk of being readmitted to a hospital within 30 days. Scholars who assessed the program determined that it could reduce rehospitalisations by 18 percent while cutting a hospital’s costs by nearly 4 percent.

The comments made by Horvitz once again highlight the need for AI companies to ensure their approach is responsible and ethical. The opportunities are endless if AI is developed properly, but it could just as easily lead to disaster if not.

Update: A previous headline ‘Microsoft has dropped some deals over AI ethical concerns’ was misconstrued as meaning the company dropped some existing deals. It has been updated to reflect Microsoft decided against some possible future partnerships over ethical concerns.

What are your thoughts on Microsoft’s approach to AI ethics?

The post Microsoft dropped some potential deals over AI ethical concerns appeared first on AI News.

In the company’s annual report for the 2017 fiscal year, AI is now mentioned on six occasions. Compared to not a single mention in the previous report, it’s clear Microsoft is now putting more of its focus on artificial intelligence.

Microsoft also indicates this in their corporate vision statement: “Our strategic vision is to compete and grow by building best-in-class platforms and productivity services for an intelligent cloud and an intelligent edge infused with AI.”

AI is only just starting to get real momentum (hence why this very publication has been launched, hello!) but it seems to be another case of where the company has slumbered, and that’s incredibly frustrating because it has much to contribute.

Microsoft is in a very advantageous position to lead in AI. First off, it has Bing. While it doesn’t have anywhere near the user numbers of Google’s search engine, it has been crawling the world’s information and building a vast knowledge graph since it launched back in 2009. This is a valuable resource and it’s only Google in the western world that’s able to compete.

When you include eastern markets, then Baidu, China’s largest search engine, is also in a similar position. Google and Baidu are considered world leaders in AI, and their search engine backgrounds are part of the reason why.

Microsoft also has the talent. Rivals including Apple and Facebook have poached some of it, but there’s good reason. Microsoft is full of it. In fact, Microsoft is the biggest supplier of AI talent to Facebook, Amazon, and Google. The company should’ve looked at why their employees were heading elsewhere long ago, I’d wager it was a lack of momentum compared to rivals.

Finally, Microsoft has the users. Whether it’s enterprises using Azure, employees, and students on Windows, or even gamers on Xbox… the company has legions of people already using their products who could benefit from AI.

Ok, it’s not like Microsoft has been completely absent from AI, the company even has a fairly capable personal assistant with Cortana which in tests is only second to Google in its ability to understand and correctly answer general questions. Compared to Siri and Alexa, Cortana is ahead by a wide margin in this regard.

But while others have been advancing their capabilities rapidly, Cortana has stagnated. Today it’s the only assistant which is unable to control smart home devices.

When the Xbox One was launched, the console was marketed as being the one entertainment center for your home. Whether you wanted games, videos, TV, or music, you were supposed to find it on the “always on” Xbox. Microsoft missed a huge opportunity to make the Xbox the center of your smart home and give people a reason to want Kinect, a device which has seemingly become the perfect metaphor for Microsoft – impressive, yet often missing the mark.

Last year, Microsoft set up a dedicated research group for artificial intelligence reporting to Harry Shum, Vice President of Artificial Intelligence & Research, which consists of 7,500 scientists, researchers, and engineers from its product teams. The group’s purpose is to speed up Microsoft’s AI development and become as active in the community as the likes of Google, IBM, and Amazon.

A new research and incubation hub called Microsoft Research AI has also been set up within its research labs unit to work on solving problems associated with AI. Part of this work is to unite research in currently disparate areas such as deep learning, machine perception, and natural language processing. By integrating these disciplines, Microsoft believes it will cause the next technology breakthroughs, such as machine reading.

This is applaudable, but not if like a lot of Microsoft’s best work it just sits in a research lab. At least to help settle the concerns about AI going rogue the company has set up the Aether ethics advisory board alongside Google, Facebook, and Amazon.

As for the importance of AI, I’d go as far as to suggest there are similarities between the AI race and two of humankind’s most defining moments. The space race, and the nuclear arms race. The former was mostly about bragging rights, while the latter was a show of deadly capability to potential opponents.

China’s strict censorship and its ‘great firewall’ is renowned but earlier this year it went a step further when the nation blocked streams of Google’s AlphaMind AI beating the country’s Go champion. While usual sports broadcasts go uninterrupted, it was clear there was a political motive behind the decision to block the competition.

As we reported earlier this week, most cybersecurity experts believe AI will be weaponised over the course of the next 12 months. Whether by individual or state-sponsored hackers, the race is on for governments to match and/or exceed those capabilities. That is the dark side of AI no-one likes to discuss, but today’s battlefield is increasingly digital, and governments will be looking to weaponise AI.

That is the importance of AI, and it’s good to see Microsoft finally putting their full weight behind it, even if it should have been much sooner.

Are you looking forward to Microsoft focusing its efforts on AI?

The post Opinion: Microsoft now considers AI a top priority – and it’s about time appeared first on AI News.