OpenAI’s warnings about dangerous AI are mostly just marketing

OpenAI has announced a new AI model, cryptically named o1, that the company says is more powerful than its predecessors — and more dangerous. But before we worry about machine apocalypse, it’s worth considering what use such warnings are.

While previous models such as GPT-4 were considered a “low” risk for public release, the new o1 model is the first to be deemed a “medium” risk by half of the criteria, according to OpenAI’s internal framework. OpenAI…