Why some people are stepping back from AI as concerns come from inside industry – explainer

Concerns are growing within the AI industry regarding the rapid deployment of advanced AI systems, with leaders like Dario Amodei (Anthropic), Sam Altman (OpenAI), and Demis Hassabis (Google DeepMind) advocating for a slower, more cautious approach. Their warnings highlight potential risks including cyberattacks, biological misuse, autonomous weapons, mass surveillance, and the spread of misinformation. Amodei, in his essay “We Must Pace the Frontier,” emphasized the need to slow the improvement of AI model capabilities, particularly concerning “frontier models” that can perform complex tasks with limited supervision, potentially leading to unintended harmful outcomes. Altman echoed this sentiment, stating, “I agree with Dario that we need to pace the frontier,” and called for independent evaluation of AI systems before release. Hassabis supported these proposals but noted that the practical details require further development. Public anxiety is also evident, with a June 2026 Pew Research Center survey indicating that 52% of Americans are more concerned than excited about AI's growing use.