With Facebook, Twitter, Google, and Amazon in the headlines seemingly every week for another controversial business or ethical decision from their past, a Benjamin Franklin quote comes to mind.
“By failing to prepare, you are preparing to fail.”
Perhaps, in an effort to avoid the setbacks, complications and legal ramifications that rapid growth fueled by technological advancements has wrought, Harry Shum, Executive Vice President of Artificial Intelligence & Research at Microsoft, warns that people need to anticipate the risks involved with the growth of artificial intelligence.
In a piece published by the MIT Technology Review and originally addressed during EmTech Digital conference, Shum warns that now is the point in the development cycle where developers and engineers can build in safeguards to help mitigate the number of unbridled risks that come with AI that’s increasingly becoming capable of being more like human beings.
“This is the point in the cycle … where we need to engineer responsibility into the very fabric of the technology,”
Even with that being said, Shum also noted that most in the industry have already been derelict in their duties to anticipate technological flaws as well as the potential negative impact of current real-world applications of AI. Items such as facial recognition come to mind when reviewing Shum’s claims as the technology has been historically bad at recognition and attribution of darker skin tones while simultaneously being weaponized by countries as surveillance and suppression tools.
Neither application of facial recognition point to a concerted effort by the industry to corral and mindfully think of the potential outcomes of the technology before they were released to the public.
While it’s one thing to caution the industry with a wagging finger, Microsoft is also taking action to help address present and future AI predicaments, by improving its own facial recognition with the arduous tasks of adding multiple skin variations to its photo recognition algorithm.
Microsoft’s facial recognition improvements are incremental in comparison to the other major action the company is taking with its AI ethics committee and its presence in the Partnership on AI. Microsoft’s vision does not only add an AI ethics review step to its own checklist of privacy reviews and accessibility audits but that the entire industry adopts the same methods.
“We are working hard to get ahead of the challenges posed by AI creation. But these are hard problems that can’t be solved with technology alone, so we really need cooperation across academia and industry. We also need to educate consumers about where the content comes from that they are seeing and using.”
However, unlike the mess that Facebook, Google, the US Congress, and the European Union lawmakers find themselves in today, Microsoft would like government agencies to step in early to help sort out the gritty details of oversight and regulation of AI technology.
Only time will tell, how the industry prepares for the tidal wave of AI-related technologies and applications, but there does appear to be sensible fear of AI getting ahead of regulation, oversight, and accountability.Further reading: Facebook, Google, Microsoft