Table of Contents
Google AI ethics has always claimed to be committed to developing artificial intelligence (AI) responsibly, in line with its own AI principles that prohibit harmful or unethical uses of the technology. But the fate of the team that has been in charge of enforcing those principles is now uncertain after its leader left and most of its members were reassigned to other divisions.
The rise and fall of RESIN, Google’s AI ethics team
The team, known as RESIN (Responsible Innovation), was founded in 2018 by Jen Gennai, who was the director of responsible innovation at Google. RESIN was responsible for reviewing internal AI projects and ensuring they complied with Google AI ethics, which were introduced after a backlash from employees and activists over a Pentagon contract that involved Google AI ethics for drone surveillance.
RESIN was part of the Office of Compliance and Integrity, within the global affairs division of Google. It conducted over 500 reviews last year, covering all of Google’s core services, including the Bard chatbot, a generative AI tool that can create text and images. RESIN also blocked the release of some AI products that could be used to create deepfakes, such as image generators and voice synthesis algorithms.
However, earlier this month, Gennai suddenly left her role as the leader of RESIN, according to four sources who spoke on the condition of anonymity. Her LinkedIn profile now lists her as a Google AI ethics and compliance adviser at Google, which the sources say indicates she will soon depart from the company altogether.
Google AI ethics team reshuffled amid uncertainty and skepticism
Around the same time, Google split RESIN into two groups, transferring 90% of its staff to the trust and safety team, which deals with abuse and misuse of Google’s services, and keeping 10% in the compliance unit. The sources say no one was laid off, but the rationale and the implications of the changes are unclear. They also say they have not been informed how the AI principles reviews will be conducted in the future.
Brian Gabriel, a Google spokesperson, declined to provide details about the restructuring of RESIN but said it was a sign of Google’s dedication to responsible AI development. He said the move “brought this particular Responsible AI team to the center of our well-established trust and safety efforts, which are baked into our product reviews and plans.” He added that it would “help us strengthen and scale our responsible innovation work across the company.”
However, some employees are skeptical about Google AI ethics ability to balance its AI ambitions with its ethical commitments, especially as the company faces increasing competition and pressure in the generative AI field, where tools like TikTok and ChatGPT are challenging Google’s dominance in search and social media.
Google AI ethics team under pressure from business and politics
Some of them also worry that RESIN’s work will be compromised by being placed under the global affairs division, which is headed by Kent Walker, Google’s chief legal officer and senior vice president of global affairs. Walker has been criticized by some activists and lawmakers for his role in lobbying and influencing policies that affect Google’s interests, such as antitrust, privacy, and content moderation.
Google has also been streamlining its operations and cutting costs in the past year, as the pandemic and the regulatory environment have affected its advertising revenue, which funds most of its AI research and development. The company has shut down some services, such as its Podcasts app, and reduced some features from Google Assistant.
Google’s CEO Sundar Pichai and other executives have maintained that they can still pursue AI innovation while being mindful of its potential risks and harms. Google joined other major AI players, such as OpenAI and Microsoft, in signing a voluntary pledge last year to evaluate the societal and national security implications of advanced AI, initiated by the White House.
Google’s AI ethics team faces uncertainty after merger with DeepMind
However, the changes in RESIN are not the only ones that have shaken up Google’s AI landscape. Last year, Google merged its UK-based AI lab DeepMind, which is behind some of the most groundbreaking AI achievements, such as AlphaGo and AlphaFold, with its main research team Google Brain, to form a new entity called Google DeepMind.
The merger aimed to unify the development of the foundational models that power generative AI tools, such as Bard and Gemini, Google’s latest AI model that can generate text, images, audio, and video.
However, the ethical oversight of Google’s most advanced AI models has also shifted as a result of the merger. According to a technical paper published last month, the responsibility for reviewing Google’s generative AI products now falls to Google DeepMind’s Responsibility and Safety Council, rather than RESIN. The paper also revealed that RESIN had influenced some of the design choices of Bard, such as limiting its use of personal pronouns to avoid anthropomorphizing it.
The future role of RESIN, which has been the primary guardian of Google AI ethics, is now unclear. The sources say they are concerned that the team will lose its influence and independence and that Google will prioritize its AI goals over its AI principles.
