In March, Google fired AI researcher Satrajit Chatterjee after he led a team of scientists to challenge a paper Google published in Nature last year. The paper in question spoke about how computers can design parts of chips much faster than a human. Chatterjee reportedly challenged some of the paper’s claims in an internal email and questioned if the technology had been tested. Google, however, stood by the paper and claimed it has been properly vetted. Google also said Chatterjee was ‘terminated with cause.’
Over the years, Google has spent billions of dollars building its Google Brain research team, which it considers key for its future. However, the controversial exits raise questions regarding the tech giant’s commitment to ethics and diversity. Also, is there a pattern to it?
Sign up for your weekly dose of what's up in emerging technology.
Let’s look at some of the top researchers who have left Google in recent times.
The most notable and most controversial on the list is the former co-lead of Google’s AI ethics team. Timnit took to Twitter to announce that she was fired; however, Google, to date, still maintains that Gebru resigned from her posts.
A paper she co-authored, ‘On theDangers of Stochastic Parrots: Can Language Models Be Too Big’, allegedly led to her termination from Google.
According to Google’s AI head, the work didn’t meet Google’s bar for publication. The paper, which builds on previously published papers, surveyed the known pitfalls of large language models like GPT-3 and listed four primary risks. Interestingly, the paper passed muster in Google’s internal review process, and objections came up much later.
Alphabet CEO Sundar Pichai agreed to investigate the matter after the company came under criticism. As a result, Google made changes to its research and diversity policies. however, the results of the investigations were not made public.
On December 2, 2021, Gebru announced the Distributed Artificial Intelligence Research Institute (DAIR). Senior research scientist Alex Hanna and software engineer Dylan Baker left Google to join DAIR earlier this year.
Margaret Mitchell, the former co-lead of the Ethics AI team at Google along with Gebru, was fired soon after the latter’s ouster.
Mitchell, like Gebru, called for more diversity in Google’s research team, and raised concerns about Google’s handling of criticism. She got the pink slip for allegedly violating the company’s code of conduct as she reportedly used automated software to scan her messages to find discriminatory treatment of Gebru.
Mitchell joined Hugging Face after Google.
Samy Bengio, a prominent member of Google Brain, also resigned after the Gebru episode. He is known for leading a large group of researchers working in ML, including adversarial settings. Samy, who served in Google for over 14 years, oversaw Gebru’s team. But, according to him, he wasn’t notified about her termination.
Samy took to Facebook to support to Gebru and said, “I stand by you, Timnit.” At present, he is a senior director of AI and Machine Learning Research at Apple.
Mustafa Suleyman co-founded AI/ML firm DeepMind in 2010 and initially raised funds from Horizons Ventures and entrepreneurs like Elon Musk.
He joined Google in 2014 when Google acquired DeepMind for USD 650 million. Then, he was Google’s vice president of product management and policy for AI. Even though he was not an AI researcher, Suleyman was instrumental in pushing DeepMind into healthcare and was vocal about not using AI for military applications.
Suleyman was accused of an aggressive management style by colleagues at DeepMind. Currently, he is the co-founder and CEO of Inflection AI.
Alex Henna, a senior research scientist at Google Brain, also resigned from Google after the expulsion of Gebru and Mitchell. She was at Google for around four years and worked on ethical AI and ML fairness.
In a Medium article, Hannah said: Many folks — especially Black women like April Curley and Timnit — have made clear just how deep the rot is in the institution. I am quitting because I’m tired” she said.
Senior research scientist Eric Jang left Google Robotics and joined Norwegian robotics company Halodi Robotics as vice president of AI. He was a t Google for nearly six years. Jang is an expert in reinforcement learning, deep learning and generative modelling.
Niki Parmar joined Google as a software engineer and later moved to the Google Brain team. She started as a research engineer, and rose to the position of a staff research scientist in 4.5 years. At Google, Parmar was involved in research related to understanding self-attention and how other inductive biases can be used for the improvement of different models across various tasks like machine translation, language modeling, and perception. She co-authored the seminal paper on Transformers — Attention Is All You Need.
Parmer is now the CTO at Adept, an AI firm she co-founded with her former colleagues at Google.
Ashish Vaswani was a member of the Google Brain research team, where his focus was on developing pure attention-based models. He also co-authored the Transformer papers. After leaving Google, he co-founded Adept, and is the Chief Scientist there.
Tatiana Shpeisman served at Google for nearly five years as a senior engineer manager, where she oversaw a team working on TensorFlow graph compiler, MLIR, and TensorFlow infrastructure for GPUs and CPUs. She is currently the compiler engineering director at Modular AI.
Augustus Odena was also a member of the Google Brain team for nearly six years. In a blog, he said: “A lot of my work has been on Program Synthesis — in which we try to get computers to program themselves. I led Brain’s work on Program Synthesis with Large Language Models, co-created Google Sheet’s SmartFill programme synthesiser, and jointly invented the Scratchpad Technique for getting Transformers to perform multi-step reasoning.” He was a senior research scientist at Google Brain team when he parted ways.