Discussions about whether to regulate AI-builds-AI must include the full range of ins and outs.
In today’s column, I examine the rising awareness among policymakers and lawmakers about potentially regulating the burgeoning use of AI to advance AI, known generally as AI-builds-AI. The crux is that there are worries that if AI is allowed to build AI, perhaps the AI being built will be beyond our control. The specter of AI existential risk looms heavily.
This topic has gotten increased attention due to the recent posting by Anthropic about their efforts to push ahead on recursive self-improvement for AI; see my coverage at the link here. Anthropic also noted that they have concerns about whether AI safety will be capable of preventing troubles and suggest that a global pause on AI-builds-AI should conscientiously be given due consideration. I discussed the floated proposition in my recent analysis at the link here.
Meanwhile, all of this has caught the ears and eyes of many policymakers and lawmakers. I provide a semblance of what new AI laws might look like to control the advent of AI-builds-AI. Whether such AI laws will be put into place is an open question. I include the ins and outs of the efficacy of those types of laws.
