Anthropic CEO Dario Amodei urged the industry to “slow the pace” of model development on Sept. 12, saying “we must slow the pace at which we improve the capabilities of AI models,” and calling for more time to let safety catch up as artificial intelligence capability rises.
Amodei published his essay and told outlets the rapid advance of artificial intelligence has outpaced society’s guardrails, writing that “Progress will still seem fast, and we must make wise use of the time” to address risks, according to The New York Times and The Guardian reporting on his remarks.
Earlier this week a former Anthropic researcher, Jacob Coxon, quit and warned that teams inside frontier labs were alarmed; the BBC reported staff were “genuinely frightened” about the technology’s direction, and NBC quoted Coxon saying he feared “systems that can improve themselves without much, or any, human input.”
Both Amodei and the departing researcher framed the concern as a gap between capability and safety. Amodei urged a deliberate pause in improving model capabilities so regulators and companies can build stronger safeguards, a position covered by TechCrunch and CBS News, which quoted his call for slowing development across the industry.
That tension has already prompted political and economic responses: U.S. lawmakers are debating stricter limits after recent high-profile warnings, and central banks have flagged macroeconomic risks from rapid AI adoption. For background on policy moves, see the bill to ban artificial superintelligence introduced by Senator Bernie Sanders and the Bank of England’s warning that AI could trigger an economic downturn.
Experts who commented to news outlets said the core worry is not a single failure but the pace at which capabilities improve. Multiple outlets reported Amodei saying industry actors should make “wise use of the time” to develop oversight and safety engineering, while former employees warned of scenarios where models might self-improve beyond immediate human control.
Industry reactions were mixed in coverage: some CEOs and researchers backed stronger safety work, and others emphasized continued innovation with added precautions. News organizations including NBC News and CNBC noted the resignation and public warnings from the ex-researcher had intensified calls for concrete governance measures.
For business readers, the key question is how companies balance commercial competition and safety investment as artificial intelligence becomes more capable. Analysts quoted in coverage say slowing capability growth could change product road maps and investment timetables, but those assessments remain tied to ongoing policy discussions and industry decisions.
Sources
- The New York Times — Amodei’s essay and the quote “We must slow the pace at which we improve the capabilities of A.I. models.”
- The Guardian — reporting on Amodei’s call to “slow the pace” and the need to let safety catch up.
- NBC News — coverage of Jacob Coxon’s resignation and the quote about “systems that can improve themselves without much, or any, human input.”
- BBC — reporting that staff were “genuinely frightened” at Anthropic, citing the ex-researcher’s warnings.
- TechCrunch — summary of Amodei’s plan and industry reaction.
