Impact of Meta and Twitter’s AI and ML layoffs

The entire ethical artificial intelligence (AI) team at Twitter, which works to improve the transparency and fairness of Twitter’s algorithms, was fired ten days ago as part of the company’s major layoffs. Rumman Chowdhury, who is renowned for her expertise in the area of applied algorithmic ethics, served as the team’s leader and was known as ML Ethics, Transparency, and Accountability.

A 50-person research team dubbed Probability that specializes in machine learning (ML) infrastructure was among the 11,000 people, or 13% of the company’s total, that Meta laid off last week. According to one researcher on the team, the Probability team consisted of 19 persons working on Bayesian modeling, 9 working on ranking and recommendations, 5 working on ML efficiency, 17 working on AI for chip design and compilers, as well as management.

Both rounds of layoffs are significant, according to analysts, as they portend a change in the landscape of even the most in-demand AI and ML expertise and a reckoning for Big Tech and enterprise businesses regarding their own responsible AI initiatives.

Meta leaders in AI and ML are fired

The Probability team at Meta is similar to an elite army tactical unit, according to Georgios Gousios, head of research at software business Endor Labs and associate professor at the Delft University of Technology in the Netherlands.

The Probability team, which Gousios worked on from October 2020 to February 2022, did work that is orthogonal to everyday software production, aiming to invent and apply new tools/methods that would make the other teams more efficient in their day-to-day work, he said. Facebook had many developers working on various parts of the tech stack and business.

He listed applications for software engineering such as tools that use ML to help engineers both write code faster with fewer bugs and debug inevitable problems more quickly. This included probabilistic programming (writing programs where variables are represented by distributions rather than single values), differentiable programming (making neural networks more efficient), and applications for neural networks.

The crew has a very high level of quality, he remarked. Many of them had decades of industrial research experience in organizations like Bell Labs or Microsoft Research, and many of them had come from years of university research. More than 60%, he believes, had doctorates.

Given how highly regarded the Probability team was, many in the AI and ML fields expressed astonishment at the layoffs.

These were some of the greatest in the field, but not as well-known as other researchers, according to Nantas Nardelli, the senior research scientist at climate tech AI business Carbon Re.

They frequently produce less impressive work, but it could end up serving as the foundation for ML goods in 5 to 10 years.

He added that their ML research is “well-applicable” to issues involving little to moderate amounts of data, substantial domain expertise, and the need to quantify uncertainty. According to him, “this skill is typically difficult to obtain, and fewer and fewer people are specializing in it these days.

Twitter ethical AI layoffs teach businesses valuable insights

The ethical AI layoffs at Twitter were not unexpected, according to Triveni Gandhi, the responsible AI lead at the data science and ML platform Dataiku.

Because of the way Twitter’s current leadership has expressed their opinions on issues of ethics, trust, and security, her immediate instinct was that they would be the first to be let go.

as a responsible AI lead, she also started to consider what the news meant for her enterprise clients: “Are they also going to start thinking, well, we don’t need this stuff? She also acknowledged that the response from the general public to the layoffs demonstrated how significant and well-respected ethical AI has grown.

She believes that other businesses are deterred from following in her footsteps by this very public decrease of that particular team, she said. she doesn’t want to instill suspicion in the minds of her AI product customers.

The announcement of Twitter’s ethical AI layoffs, she continued, is inspiring a “feeling of commitment” among her clients. They are claiming that they can do better and that they will empower [responsible AI teams] to begin putting ideas into practice.

Source link