AINewsnow

AI Safety Push: Anthropic and OpenAI CEOs Call for Slowing Development

2026-09-15

AI Safety Push: Anthropic and OpenAI CEOs Call for Slowing Development

Major AI leaders are pivoting toward caution, with Anthropic and OpenAI CEOs urging a slowdown in development to prioritize safety. This shift coincides with significant product launches from Apple and DeepSeek, alongside geopolitical tensions over AI governance.

  1. Anthropic CEO Dario Amodei Calls for AI Development Slowdown Anthropic CEO Dario Amodei stated it is time to slow down AI development and announced that third-party evaluators like METR will gain access to its models. This move aims to ensure adherence to safety practices and commitments as outlined in his recent essay.
    Why it matters: This marks a significant shift in strategy for leading AI labs, prioritizing safety over speed and inviting external scrutiny.
  2. OpenAI Delays IPO Citing Safety Concerns OpenAI CEO Sam Altman confirmed in a Fortune interview that the company will not go public this year due to necessary safety work. This decision comes amid rising fears about doomsday AI scenarios and industry-wide safety lags.
    Why it matters: Delaying the IPO signals that safety concerns are now a primary driver of corporate strategy, potentially impacting investor expectations.
  3. Apple Releases iOS 27 and macOS Golden Gate with Siri AI Apple has officially released iOS 27 and macOS Golden Gate, integrating the new Siri AI into its operating systems. Users in the EU and China are still experiencing kinks with the new features.
    Why it matters: The widespread rollout of Siri AI brings advanced generative capabilities to hundreds of millions of consumer devices.
  4. DeepSeek Unveils v4.1-Flash with Novel Encoder-Decoder Architecture DeepSeek released v4.1-Flash, featuring a 763B-P8B-D16B novel causal Encoder–Decoder architecture with vision capabilities. The community has noted that the significant architectural changes suggest this should have been labeled as version 5.
    Why it matters: This release demonstrates continued rapid innovation in open-weight models, challenging existing benchmarks with new architectural approaches.
  5. Anthropic and OpenAI CEOs Jointly Urge Caution in AI Development Dario Amodei and Sam Altman have both publicly urged caution, highlighting fears that the industry is racing too fast to build powerful AI while safety measures lag behind. Their statements reflect a growing consensus among top lab leaders about the need for a pause.
    Why it matters: Unified calls for caution from the heads of the two most prominent AI labs could influence regulatory frameworks and public perception.
  6. Anthropic Whistleblower Advocates for Self-Regulation Room Jacob Coxon, a former Anthropic employee, stated that AI companies need room for self-regulation while sparking concern about the pace of advancement at labs like Anthropic, OpenAI, and Google DeepMind. His comments have fueled widespread debate about internal safety cultures.
    Why it matters: Whistleblower insights provide a rare look into the internal tensions between competitive pressure and safety protocols at leading AI firms.
  7. China Proposes BRICS 'Open-Source AI Zone' President Xi Jinping announced that China will lead the creation of an 'open-source zone for artificial intelligence' for BRICS countries during a summit in New Delhi. This initiative aims to foster collaborative AI development among the 11-nation group.
    Why it matters: This proposal highlights China's strategy to build alternative AI ecosystems and strengthen technological ties with emerging economies.
  8. Amazon Bedrock AgentCore Adds OAuth Consent Management Amazon Bedrock AgentCore Identity now offers a Consent portal, a managed web experience, and a session binding endpoint for AgentCore Gateway. The update simplifies provisioning portals and configuring third-party targets like GitHub and Slack.
    Why it matters: Enhanced security features for AI agents are critical as enterprises increasingly deploy autonomous systems requiring user data access.
  9. Experts Weigh In on AI Existential Risk Claims University of Southampton professor Thomas Dekeyser discussed concerns about AI becoming fatally out of control in response to claims by a former Anthropic employee. The discussion highlights the ongoing debate about the severity of AI safety risks.
    Why it matters: Academic scrutiny of existential risk claims helps ground public discourse in evidence-based analysis rather than speculation.
  10. China Dismisses AI 'Fearmongering' Amid US-China Tensions Beijing dismissed AI 'fearmongering' after Anthropic CEO Dario Amodei called for the US to impede China’s progress in AI. China's spy chief warned of threats to Communist party rule, framing the issue as a geopolitical security matter.
    Why it matters: The rhetorical clash underscores the deepening geopolitical divide over AI governance and national security implications.