The Self-Regulation Paradox
Speaker Mike Johnson's position going into Tuesday's White House meeting was unambiguous: 'I hope it's voluntary' [1]. His underlying theory is that competitive pressure, not statute, will keep AI companies honest - he says the industry wants to reassure 'investors, consumers, the American public that they're going to do this in a very safe way' [2]. But Johnson has also acknowledged, elsewhere, that the industry leaders he consults - including Anthropic's Dario Amodei, OpenAI's Sam Altman, and Elon Musk - all say guardrails are needed but have very different ideas on what that looks like, meaning they don't themselves even have the answer as of today. That is the paradox at the center of the voluntary-guardrails plan: the entities being trusted to self-police cannot agree internally on what policing even means, and Johnson has offered no mechanism for reconciling those differences beyond calling, in a separate interview, for platform providers to sit down together and hash out shared standards - a coordination step distinct from the one-on-one conversations he says he's already had with each of them.



