Top VC firms pledge responsible strategy on AI startups

Coalition says it wants to ensure it funds firms with strong guardrails.

Nearly three dozen venture capital firms have signed a set of voluntary commitments, with feedback from the Biden administration, for how the many startups they back should develop artificial intelligence technology responsibly.

Responsible Innovation Labs, a nonprofit coalition of investors and tech executives, is set to release details on Tuesday of the new guidelines, which include pledges to “secure organizational buy-in” within startups around responsible AI, “forecast AI risks and benefits,” and “audit and test to ensure product safety.” The group consulted with the U.S. Commerce Department, as well as academics and civil society groups, to create the voluntary agreement, and will discuss it at a roundtable in San Francisco with Commerce Secretary Gina Raimondo on Tuesday.

Under President Joe Biden’s recent executive order on artificial intelligence, the Commerce Department has been tasked with setting evaluation standards for AI companies, but these mandatory evaluations will only apply to firms that build the most powerful AI models. The new voluntary guidelines, while open-ended, are part of an effort to enact some guardrails for potentially thousands of startups across the AI industry.

The 35 firms that have initially signed on to the commitments include General Catalyst, Felicis Ventures, Bain Capital, IVP, Insight Partners and Lux Capital. In addition to the guidelines, RIL is also releasing a more detailed “protocol” — or how-to guide — for startups and investors trying to implement these principles, including suggestions on how to structure responsible AI teams, run safety tests and engage with outside stakeholders.

“This brings a kind of additional level of hygiene, if you will, into how you manage investments,” Ganesh Bell, managing director of Insight Partners, told Bloomberg News in an interview. “We know from investing in infrastructure and AI companies that it’s important to actually build trust. It’s important to build safety.”

In a statement, Raimondo said the announcement reflects Biden’s “all-hands-on-deck approach to ensure we are leaving no stone unturned to harness the power of AI for good while protecting people from its risks.” The voluntary agreements, she said, “demonstrate important leadership from the private sector.”

As these standards aren’t binding, it will be up to VCs and startups to self-enforce. Gaurab Bansal, executive director of RIL, said he’s hopeful that firms will uphold the new principles. “We think that founders will be looking to the signatories to make sure that they’re following through on their VC commitments, and vice versa,” he said.