Nvidia CEO Jensen Huang said at the Dreamforce event that new regulations are not needed for artificial intelligence safety and that responsibility should lie with the companies developing the products. Viewing safety as an engineering problem, Huang argued that companies should not release products unless they are confident in their functionality and safety.
Anthropic CEO Dario Amodei, meanwhile, called for greater investment in safety, transparency, independent audits and common industry standards. Cohere CEO Aidan Gomez said an antitrust exemption that would allow companies to coordinate on safety could create a cartel that restricts competition among major artificial intelligence companies.
Why it matters
This divergence of views highlights the fundamental conflict between self-regulation and external oversight in artificial intelligence safety. Leaving responsibility solely to developer companies leaves open the questions of how safety will be verified and how accountability will be ensured after a product is released. While calls for transparency, independent audits and common standards aim to close this gap, the risk that coordination among companies could restrict competition creates a separate problem. The debate requires regulators to consider safety obligations and competition law together. For the industry, the key dividing line is whether safety will be ensured through companies’ own engineering processes or through common and auditable rules.
Background
Nvidia is not a new name in the FikirPilot archive: over the past 90 days, we have published 32 news stories mentioning the name, the most recent dated September 17, 2026.