Tech Giants Will Explore SAFA Leadership for AI Safety

Google, OpenAI, and Anthropic are considering Sriram Krishnan to lead the new industry-led standards organization.

Updated on Sept. 25, 2026 in Artificial Intelligence

Tech Giants Will Explore SAFA Leadership for AI Safety

Live Poll

Do you trust private industry groups to set safety standards for artificial intelligence?

Major AI developers including Google, OpenAI, and Anthropic are evaluating Sriram Krishnan to serve as CEO of the proposed Standards Authority for Frontier AI (SAFA). The organization is slated for a potential launch by late 2026 or early 2027.

Why it matters

Proponents argue that current government oversight bodies lack the necessary funding and technical expertise to properly evaluate advanced AI models. This industry-led initiative aims to bridge that gap by establishing testable benchmarks for safety.

The proposed SAFA body intends to transform broad public safety pledges into concrete, testable benchmarks for developers. It would provide the technical infrastructure needed to evaluate powerful models before they are released commercially.

The players

Sriram Krishnan

He is a former White House AI policy adviser and current general partner at Andreessen Horowitz who has held roles at Microsoft and Twitter.

Demis Hassabis

He is a prominent figure in the AI industry who proposed the creation of a US-based standards body in July 2026.

Google

This major technology company is one of the three primary firms collaborating to establish the industry-led safety organization.

OpenAI

This AI research and deployment company is a lead planner in the effort to create new testing benchmarks for frontier models.

Anthropic

This artificial intelligence safety and research company is working with industry peers to formalize independent safety standards.

The details

The proposed organization functions as an independent, industry-led body rather than a government regulator. Sriram Krishnan, a general partner at Andreessen Horowitz and former White House AI policy adviser, is the primary candidate currently being considered to spearhead the effort.

January 2025 to June 2026: Sriram Krishnan served as a White House AI policy adviser.
July 2026: Demis Hassabis formally proposed the US-based AI standards body.
Late 2026 or early 2027: Companies aim to launch the new organization.

Roadmap

The push for industry-led oversight reflects a pivot away from strictly governmental regulation toward self-governed technical standards. This evolution suggests that major developers are seeking to consolidate safety protocols to ensure standardized testing across all frontier AI models.

The creation of this body could standardize safety benchmarks, potentially leading to more reliable and secure AI-powered consumer tools. Users may eventually see higher quality assurance markers on software products before they hit the market.

The takeaway

The move toward independent safety authorities signifies a maturation in how AI developers balance rapid innovation with public safety. Companies seeking to implement industry-led standards often set a baseline expectation that individual consumers and regulators will eventually demand.