Former OpenAI Board Members Call for Government Regulation in Light of CEO Controversy
Former OpenAI board members, Helen Toner and Tasha McCauley, have written an Op-Ed in The Economist advocating for greater government regulation of the company. Their call for regulation comes in response to concerns over CEO Sam Altman’s leadership and alleged patterns of behavior that undermined the board’s oversight and internal safety protocols.
Toner and McCauley joined OpenAI with optimism, drawn to the company’s nonprofit status and seemingly altruistic motivations. However, they soon became disillusioned with Altman’s actions and the company’s direction. They claim that multiple senior leaders raised concerns about Altman’s “toxic culture of lying” and psychological abuse. Additionally, they express worry over Altman’s reinstatement to the board and the departure of senior safety-focused talent, which they believe jeopardizes the experiment in self-governance.
In their Op-Ed, Toner and McCauley argue that self-regulation is insufficient for ensuring the safe and ethical development of advanced AI systems. They emphasize the need for external oversight, especially in the face of immense profit incentives. They contend that governments must play an active role in regulating AI companies like OpenAI.
The former board members point to recent developments in the AI regulation landscape to support their argument. They highlight the EU’s billion-dollar warning to Microsoft for failing to disclose potential risks of their AI-powered technologies. They also reference a UK AI Safety Institute report that uncovered vulnerabilities in public Large Language Models (LLMs) when exposed to malicious prompts.
OpenAI itself has faced significant scrutiny recently due to high-profile resignations and concerns about its future. Co-founder Ilya Sutskever and his co-leader Jan Leike left the company, prompting OpenAI to disband its in-house safety team. Leike expressed concerns about the company’s prioritization of “shiny products” over safety culture and processes.
Altman has also faced criticism for OpenAI’s off-boarding policy, which requires departing employees to sign NDAs preventing them from making negative statements about the company. In response to the controversy, Altman and co-founder Greg Brockman emphasized the importance of elevating safety work and collaborating with governments and stakeholders.
Toner and McCauley argue that the traditionally light-touch approach to internet regulation is insufficient for the unique challenges posed by AI. They believe that government regulation is necessary to ensure the safe and ethical development of AI systems. Their call for greater oversight aligns with the growing recognition that self-regulation alone may not be enough to mitigate the risks associated with advanced AI technologies.