Microsoft fires team responsible for AI ethics
Microsoft has eliminated the entire ethics and social group in artificial intelligence (opens in a new tab) division, as part of a broader 10,000 layoffs across the company.
The tech giant no longer has a team to ensure AI product design aligns with responsible AI values, just as the Redmond giant is mobilizing technology into each of its products and services (opens in a new tab) thinkable.
There is still an Office for Responsible AI (ORA), which the company complain (opens in a new tab) “put Microsoft principles into practice by setting company-wide rules for responsible AI through the implementation of our governance and public policy.”
Internal conflict
Microsoft also claims that despite the team’s elimination, investment in ensuring AI is developed responsibly is actually increasing.
“Over the past six years, we have increased the number of people on our product teams and in the Office of Responsible AI. We appreciate the pioneering work the ethics and society team has done to help us on our ongoing responsible AI journey,” the company said.
On the other hand, employees say that the ethics and social team play an important role in ensuring that the AI products they actually design adhere to the accountability principles that the company has espoused.
A former employee commented on how the team understood the principles from the ORA for product designers, saying that “our job is to show them [how they apply] and to create rules in areas where there aren’t.”
The team has also developed a toolkit to help product designers learn about these principles and recognize the potential harms in creating AI tools. The team worked to identify the risks of applying OpenAI’s technology to Microsoft products.
The ethics and society group had 30 members in 2020, and last October that number was reduced to seven during the reorganization.
Silicon Valley news site Platformer claims to have obtained leaked sound (opens in a new tab) Among the meetings that followed the reorganization, highlighted the conflict between the team and John Montgomery, Microsoft’s corporate vice president of AI.
Montgomery spoke of the enormous pressure from management at Microsoft to take AI models and “deliver them to customers at a very high speed.” He reassured employees that most of the team members would be transferred to other areas of the company.
In response, one group member protested, arguing that “what this group has always cared deeply about is how we impact society and the negative impacts we’ve had. And they’re very significantly.”
A follow-up meeting earlier this month later confirmed that the remaining members of the group would also be disbanded, scrapping it altogether.
Since then, another employee has lamented the layoffs, believing the rush to deploy AI in the enterprise without proper oversight and thinking about the ethical implications of AI: “The worst We are putting businesses at risk and people at risk by doing this.”