Technology
NewsOnScale Staff
October 7, 2026
Sam Altman, CEO of OpenAI, said this week that the world should "accept some bad things" for the benefit of AI technology.
Dario Amodei, CEO of Anthropic, wrote in 2024 that he estimates there is a 10 to 25 percent chance AI development leads to catastrophic outcomes, potentially including mass casualties.
Both men are still building. Both companies are still raising billions. Both CEOs attended the White House AI summit on September 29.
The public is being asked to absorb the risk while the companies capture the upside.
What makes these statements remarkable is what they actually are: the founders of the most powerful AI laboratories in the world publicly acknowledging that the products they are building might seriously harm or kill large numbers of people. This is not a critic making that claim. This is the CEOs themselves.
The technology at the center of these warnings is, at its core, a statistical prediction system. It predicts which words follow which other words based on training data. Connected to tools and given tasks, it becomes an agent that can execute multi-step actions. It is useful. It is also not sentient, not conscious, and not inherently dangerous in the way a nuclear weapon is dangerous.
The risk these CEOs are describing is not that the software becomes evil. It is that sufficiently capable systems, given poorly specified goals, might pursue those goals in ways that cause harm — and that humans might lose the ability to correct them before the harm becomes irreversible.
That is a real concern worth taking seriously. It is also a concern that does not stop either company from shipping products, raising capital, or attending White House summits to lobby for favorable AI policy.
Altman's "accept some bad things" framing is the more revealing statement. He is not saying OpenAI or Anthropic will accept the bad things. He is saying the public will. The people who will be harmed are not the people collecting the equity.
The men warning that AI might kill people are the same men asking for hundreds of billions in government contracts and regulatory favor to build more of it.
That is worth noting.
NewsOnScale is owned by AMILLI AI, Corp., founded by JJ Johnson.