Anthropic

Anthropic's IPO Has A Safety Problem, And NYC Isn't Backing Down

New York City Comptroller Mark Levine criticized Anthropic’s planned IPO debut this week, calling it "disappointing" following previous attempts to get the AI company to delay its public offering over safety concerns.

"We’re not done fighting yet. We intend to be vocal and communicate with management and rally with like-minded institutional investors around the country and push for transparency and good safety protocols," Levin told Bloomberg Surveillance. 

Levine, who oversees New York City’s $325 billion public pension funds, previously called for Anthropic to delay its IPO. He argued the company should prioritize safety measures and greater transparency before pursuing an IPO.

The letter raised questions about how public-market investors would assess Anthropic’s business as the company itself warns about the potential dangers of increasingly capable AI systems.

"This is cognitive dissonance right now because we’ve had months of serious safety incidents. We have a company that acknowledges their technology could lead to catastrophic risks. It’s a contradiction that the company is going public right now," Levine told Bloomberg.

Levine also takes issue with the company’s governance structure, which gives founders and an independent trust disproportionate control while limiting shareholder influence over a company developing potentially dangerous AI technology.

AI Exposure is Already Widespread

The comptroller added that AI exposure is already widespread across the pension portfolio, spanning private equity, private credit, infrastructure and emerging markets. The comptroller’s office is now discussing AI-related investment risks with public pension funds across the country.

Earlier this week, the California Public Employees Retirement System (CalPERS) stated it is taking a closer look at the potential risks stemming from artificial intelligence, as the board expresses concerns regarding the industry’s effects on both workers and the environment.

Staff told board members at a committee hearing that cybersecurity threats and questions about AI model reliability are becoming an important part of the pension fund’s due diligence process.

According to data from Equable Institute, a nonprofit that studies public retirement systems, CalPERS has approximately $65 billion held in AI-related companies.

Meanwhile, Anthropic rival OpenAI’s adviser Dean Ball recently penned an essay stating that AI could eventually push humans out of positions of economic and political influence, not through a machine uprising, but through the market forces driving AI adoption.

Ball warned that society could gradually accept a future in which increasingly capable AI systems run businesses, allocate capital and influence major institutions, leaving people with less control over the direction of civilization.

Ball’s essay reflects his personal views and does not necessarily represent the views of OpenAI or its employees, the company noted alongside the publication.

Photo: RixAiArt / Shutterstock

Market News and Data brought to you by Benzinga APIs
Share this article