A study by Dutch peace organization PAX marks Microsoft as one of the tech companies most likely to be creating killer robots. The organization graded a total of fifty tech companies and their role in AI, pattern recognition, autonomous vehicles, and more.
It looked at whether the companies were developing tech relevant to autonomous weapons, previous work on military weapons, and their prior commitments on the matter. It then contacted the companies for comment, with some interesting findings.
Microsoft and Amazon were the two major tech companies marked as a high concern. Others included Oracle, EarthCube, Citadel Defense, and Shield AI. Unlike Google, Microsoft does not outright state that won’t use its tech for weaponry.
Close U.S. Military Ties
Indeed, it’s providing the U.S. military with its HoloLens 2 headset for use in warfare. INn a $480 contract, it’s proving augmented reality systems that tie in thermal vision, compass readings, virtual maps, and a gun reticle.
Director of Microsoft Research Eric Hovitz told PAX that “Microsoft makes a priority of the responsible development and use of AI technologies. Microsoft’s cross-company Aether committee takes sensitive uses of AI technologies very seriously and deliberates carefully in making recommendations to our company’s leadership team about controls and guidelines, including the critical need for human oversight and human-in-the-loop on high-stakes, sensitive AI technologies.”
However, PAX says this statement can’t be taken as the company’s official position on autonomous weapons. In a blog post last year, Microsoft President Brad Smith said the firm would continue to sell software to the military despite complaints about HoloLens 2 and its bid for the lucrative JEDI contract.
“We believe in the strong defense of the United States and we want the people who defend it to have access to the nation’s best technology, including from Microsoft,” Smith said.
No Official Policy
Microsoft has not denied that it’s working on autonomous weapons and is yet to announce an official policy regarding their creation. Though some may picture an army of human-like robots, a bigger threat could be AI-controlled drones, artillery, or cyber weapons.
“Why are companies like Microsoft and Amazon not denying that they’re currently developing these highly controversial weapons, which could decide to kill people without direct human involvement?,” queried Frank Slijper, the lead author on the report.
A Teamblind survey from earlier in the year suggests that 58% of Microsoft employees are willing to create AI/facial recognition software for use by law enforcement. It ranked the highest in this of any tech company, followed by Amazon at 56.45%.
However, Microsoft employees also ranked as the most concerned about the misuse of AI. 76% of those surveyed noted concern, versus Amazon’s 72% and Google’s 69%.