AI giants pursue self-regulation as safety fears mount

In recent weeks, escalating concerns over the safety of artificial intelligence (AI) have emerged, prompting major AI companies to explore the establishment of a self-regulatory body. OpenAI, the creator of ChatGPT, announced on Tuesday that it is collaborating with Anthropic and Google DeepMind to develop such a body, inspired by a proposal from Google DeepMind's Demis Hassabis. Hassabis has previously advocated for the creation of a US organization to test powerful AI systems prior to their public release.

Funding for the proposed body would likely come primarily from the industry, as it would require substantial resources to attract top technical talent and provide the necessary computational power for extensive testing. OpenAI CEO Sam Altman expressed support for industry-wide collaboration, acknowledging public fears that competition between companies and countries might lead to safety being neglected.

Altman emphasized the industry's desire to coordinate and ensure adequate time for safe development, particularly in response to fears that some industry players might fail to act responsibly. However, some tech figures, including the CEO of Canadian AI company Cohere, have accused dominant American AI labs of forming a cartel under the guise of safety. Cohere's co-founder and CEO, Aidan Gomez, argued that the true dispute lies in who writes the guardrails, participates, and whose interests the rules protect.

Growing concerns about AI safety have been fueled by warnings from former employees of major AI companies, who warn of the industry's risks. Last week, researcher Jacob Coxon resigned from Anthropic, expressing concern that AI labs are "gambling with our lives." Former Google DeepMind researcher Bilal Chughtai echoed similar sentiments, stating that AI has the potential to "kill us all" and urging a slowdown in AI development.

Dario Amodei, CEO of Anthropic, has also called for a slowdown in AI development, a sentiment shared by OpenAI CEO Sam Altman, Elon Musk, and Hassabis. US Senator Bernie Sanders has criticized tech leaders for failing to honor pledges to slow down development and establish safety standards, asserting that binding international safety rules are necessary rather than voluntary standards from the industry.

Altman warned that AI accidents, both minor and catastrophic, may be unavoidable but emphasized the importance of transparent reporting and learning from mistakes quickly to minimize larger incidents. US President Donald Trump dismissed AI safety concerns as "hoaxes," while Nvidia CEO Jensen Huang dismissed the warnings as part of a coordinated campaign.