Ex-Rep Alarmed by Briefing on AI's Potential Perils
Former Rep. Will Hurd was "freaked out" by a briefing on AI system GPT4 while serving on OpenAI's board, calling it a first step toward artificial general intelligence (AGI). He argues for governance guardrails on AGI given the potential for both good and catastrophic consequences.
Summary
- Former Rep. Will Hurd served on OpenAI's board for two years before stepping down to run for president.
- At one point, Hurd received a briefing on GPT4 that deeply concerned him about the progress toward AGI.
- Hurd says AGI could enable solutions to global problems but also lead to irreversible consequences like those of nuclear war if unchecked.
- He calls recent leadership turmoil at OpenAI a "call to action" to ensure AGI is governed properly.
- Hurd argues AI should be accountable to laws, compensate creators, and require permits for powerful systems.
- A permitting process would ensure powerful AI operates by agreed upon safe standards, similar to permits for nuclear plants.
- Governance of rapidly advancing AI capabilities can't be left to a few people given the stakes involved.
- There is a need to put guardrails in place to make AGI "a force for good rather than the harbinger of catastrophic consequences."