Openai's Altman warns AI accidents are inevitable, calls for transparent reporting

On September 16, 2026, Sam Altman, CEO of OpenAI, warned that some AI accidents are unavoidable but emphasized the importance of learning from them. Speaking at Salesforce's Dreamforce conference, Altman urged tech companies to adopt a culture of transparency, drawing parallels to the aviation industry's approach with the FAA and NTSB. He acknowledged public fears about AI's potential risks, stating that the world's fear is justified due to the possibility of AI losing control and the concentration of power in a few labs. Altman argued that AI development should prioritize alignment, safety, and monitoring over rapid model advancement. Meanwhile, Trump's administration has shown a lack of regulatory interest in AI, with David Sacks, former White House AI advisor, advocating for unfettered development. Recent concerns about AI safety have intensified, with former Anthropic researcher Jacob Coxon warning that companies are rushing toward self-improving superintelligence. Other researchers have echoed these fears, noting that advanced AI could lower the barriers for catastrophic threats. Altman's comments have sparked criticism from AI safety advocates, who argue that labs acknowledge risks but continue to develop more powerful models. He stressed the need for transparency and rapid learning from incidents to prevent escalation, even as regulatory stances remain divided.