Update on AI Ethics: Carnage at Twitter; Fired Ethicists Spreading Out, Reemerging
Elimination of the AI ethics team at Twitter continues a trend of challenges big tech faces in trying to address thorny issues; experienced ethicists now joining like-minded organizations
By John P. Desmond, Editor, AI in Busines
The AI ethics team at Twitter was a casualty of Elon Musk’s swashbuckling entrance into big tech social media ownership and management.
AI ethics is proving to be a risky profession, with exits at Google setting the pace before the latest carnage at Twitter.
Is AI ethical? It remains to be seen, it seems.
Here is a rundown:
Musk dissolved a team well-respected for its exploratory work in ethical AI and algorithmic transparency, according to an account in TechCrunch. The team was directed by Rumman Chowdhury, whose position was eliminated along with the team’s engineers and others, according to the account.
Joan Deitchman, senior engineering manager, ML ethics, transparency and accountability at Twitter, posted a message on Twitter:
“Yep, the team is gone. The team that was researching and pushing for algorithmic transparency and algorithmic choice. The team that was studying algorithmic amplification. The team that was inventing and building ethical AI tooling and methodologies. All that is gone.”
Google AI ethicist Timnit Gebru separated from the company in December 2020 and started her own organization, the Distributed AI Research Institute (DAIR) to continue her AI research. DAIR later received $3.7 million in funding from a range of foundations. Her research at Google into risks associated with large language models ran into headwinds.
Margaret Mitchell, Gebru’s manager at Google, followed her out soon after and is now chief ethics scientist at Hugging Face, an AI community looking to “democratize” good machine learning. Hugging Face offers ready-to-use machine learning models, proving popular with developers. The company announced in May that it had raised $100 million in venture capital
Satrajit Chatterjee, a researcher in Google’s Brain group, was let go in May after being critical of a research paper on computers designing computer chips. Chatterjee disputed part of the paper that presented a method for automatically generating parts of a computer chip more efficiently than humans. Google told his team that it could not publish a rebuttal of some of the claims made in the paper. Some researchers who worked on the rebuttal told The New York Times that Google broke its own AI principles by rejecting the paper.
Google ethical AI researcher Alex Hanna published a resignation letter earlier this year describing a "whiteness problem" within Google and other tech companies, according to an account in Protocol. Hanna followed software engineer Dylan Baker in joining Gebru's DAIR, which launched late last year and will soon celebrate its first anniversary with an online AI event.
Work of AI Ethicists Doubted by Some
Some in the AI field are skeptical that the work of AI ethicists will have practical value.
“People working in the AI field are mostly engineers. They’re not really open to humanities,” stated Emmanuel Goffi, an AI ethicist and founder of the Global AI Ethics Institute, a think tank headquartered in Paris, in a recent account in MIT Technology Review.
Companies tend to want a quick technical fix, and the AI ethics field tries to fix deeply ingrained issues such as racism, sexism and discrimination that exist with the large data sets used to train AI models. That could take some time, and may not lend itself to executives who want someone to “explain to them how to be ethical through a PowerPoint with three slides and four bullet points,” Goffi stated.
The ex-Google AI ethicists are now fanning out through the industry. Hugging Face, which started out five years ago offering an iPhone chat app for teens, now finds itself as a central depot for ready-to-use machine learning models, according to a recent account in Forbes, which included the company on its 2022 AI 50 list. The developer community has built more than 100,000 machine learning models on the platform.
“Machine learning is becoming the new way to build technology, replacing software,” stated Clément Delangue, cofounder and CEO of Hugging Face, which is named after the emoji that looks like a smiling face with jazz hands. “The old school of building technology was writing a million lines of code. Machine learning is starting to do that, but much better and much faster.”
Hugging Face has made a commitment to further ethical AI. “Originating as an open-source company, Hugging Face was founded on some key ethical values in tech: collaboration, responsibility, and transparency,” stated Margaret Mitchell, in the first edition of Ethics and Society Newsletter published on the company’s blog. “How to operationalize ethics in AI is an open research area,” she states
While theory and scholarship on applied ethics and AI have existed for decades, applied practices for ethics within AI development have only begun to emerge in the last 10 years, Mitchell noted. She stated, “For those of us interested in advancing ethics-informed AI, joining a machine learning company founded in part on ethical principles, just as it begins to grow, and just as people across the world are beginning to grapple with ethical AI issues, is an opportunity to fundamentally shape what the AI of the future looks like.”
The “Ethics and Society regulars” at Hugging Face will keep this work going, and they have invited developers outside the company to participate and provide feedback, evaluate models and analyze code for bias. In addition, “We are also developing new open and responsible AI licensing, a modern form of licensing that directly addresses the harms that AI systems can create.” she stated.
And in this way, a generation of AI ethicists is at work, many of them having learned from experiences doing AI ethical work in big tech.
Read the source articles and information in TechCrunch, in Protocol, in MIT Technology Review, in Forbes and in the Ethics and Society Newsletter on the blog of Hugging Face.
(Write to the editor here.)