The news
David Sacks stated that leading AI companies should deliver safe technology on their own terms rather than request relief from antitrust rules. He named Anthropic PBC and OpenAI specifically during a Bloomberg TV interview on September 14. Sacks argued that these firms can manage safety and set their own development pace without government involvement.
The comments came directly from his role advising the administration on science and technology policy. No other companies received mention in the interview. The position rules out any trade of regulatory forbearance for voluntary safety pledges.
Context
Sacks serves as co-chair of the President’s Council of Advisors on Science and Technology. His comments come at a time when several AI developers face scrutiny over market concentration and product risks. The prior expectation in parts of the industry had been that safety concerns might justify temporary exemptions from standard competition enforcement. Sacks rejected that approach outright.
The interview occurred against ongoing public discussion of how advanced AI systems should be developed and released. Several labs had previously floated the idea that safety work could warrant lighter antitrust oversight. Sacks closed that door in plain terms during the Bloomberg appearance.
Details
Sacks told Bloomberg that top labs “should make safe products — and pace their own progress if necessary — without government intervention.” He did not propose new legislation or enforcement changes. Instead he placed responsibility for risk mitigation directly on the companies. The interview contained no discussion of timelines, specific safety benchmarks, or enforcement mechanisms. Anthropic and OpenAI were the only two organizations cited by name.
The phrasing left no room for negotiated carve-outs. Sacks framed internal safety discipline as a normal obligation of any company shipping advanced systems. No additional policy tools or review processes were described.
Reactions and counterpoints
No immediate public response from Anthropic or OpenAI appeared in the source reporting. The interview transcript released by Bloomberg contained no back-and-forth on whether the companies had previously sought such exemptions. Observers will watch for follow-up statements from either firm in the coming days.
Why it matters
Sacks’ position removes any expectation that regulators will trade antitrust forbearance for voluntary safety commitments. Companies that had considered lobbying for such a bargain now know the administration’s science-advisory channel is not open to that framing. The statement also narrows the public rationale for self-imposed slowdowns to internal decisions rather than negotiated policy outcomes. Developers must therefore justify any reduction in training runs or release cadence on technical or commercial grounds alone.
This leaves safety work as an ordinary cost of doing business rather than a bargaining chip in regulatory talks. Firms that fail to demonstrate credible internal controls will face the same competitive enforcement environment as any other concentrated market participant. The message is consistent with a broader stance that treats advanced AI as a standard industrial activity whose risks are best addressed through product discipline, not special exemptions.
The practical effect is to shift the burden entirely onto company engineering and governance teams. Boards and executives at the named labs now operate without the possibility of citing external policy negotiations as cover for slower roadmaps. Investors and competitors alike can price in that reality when evaluating capital plans or partnership terms. Over time the approach may encourage more transparent publication of internal safety processes, since external relief is off the table.
---
Sources:
{"word_count": 612, "sources_used": 1}
No comments yet