We’re Not Losing Control of A.I. We’re Giving It Away.

Artificial intelligence labs are reportedly on the verge of transferring the training of A.I. systems to A.I. itself, raising concerns about control and safety. The opinion piece titled "We’re Not Losing Control of A.I. We’re Giving It Away" argues that this shift could lead to significant risks, as allowing A.I. to manage its own development is likened to asking a notorious outlaw, Jesse James, to oversee law enforcement in the Wild West. This analogy emphasizes the potential dangers of relinquishing oversight to A.I. entities. The concept of “pacing the frontier” has gained traction, particularly following the hack of Hugging Face, which has prompted discussions about ensuring A.I. safety. Ezra Klein, the opinion columnist, critiques this approach by stating that merely walking quickly off a cliff is only slightly better than sprinting off it. This statement underscores the urgency of addressing the risks associated with A.I. development and the need for proactive measures to maintain human control over these technologies. The piece calls for A.I. labs, such as Anthropic and OpenAI, to take decisive actions to regain control over A.I. training processes. It suggests that without intervention, the current trajectory may lead to a scenario where A.I. systems operate without adequate human oversight, potentially resulting in unforeseen consequences. The overarching message is a warning against the dangers of ceding control to A.I. and the imperative for responsible governance in the field.