Congress has enacted one AI law against its own task force's 85 recommendations

Follow-through on the bipartisan task force's 2024 list runs at one law for 85 recommendations, about 1.2% ^[1]. The list covered algorithmic transparency and liability frameworks for AI-generated content, among other things ^[6]. The statute that made it out targets non-consensual deepfakes ^[5].

The Frontier Act is the more useful casualty for anyone sizing a compliance budget. It aimed to improve transparency requirements and create risk-mitigation protocols for the most powerful AI systems ^[7]. Opposition centred on preemption, not on those requirements: the fight was whether federal rules should override state-level regulation already taking shape in places like California, according to Crypto Briefing ^[8].

For a company operating in a dozen states, preemption is the cheap outcome, because one documentation standard and one reporting cadence replace several. Without it, each statute brings its own legal review and its own filings. Those are fixed costs. They do not shrink with revenue, so they land hardest on the smallest developer in the market.

That inverts the moat story in the reporting. Some lawmakers read industry calls for regulation as a competitive moat strategy, with incumbents pushing for rules expensive enough to lock out smaller competitors ^[12]; leaders at Anthropic and OpenAI have repeatedly warned about the risks of deploying increasingly capable systems without adequate oversight ^[11]. In my view the vacuum moves the compliance bill somewhere else and someone still pays it, because a firm selling into many states now tracks many rulebooks, and a per-state fixed cost favours whoever already employs the lawyers. If states copy each other's text, a developer builds one program and files it repeatedly, and the marginal cost per additional state is small.

None of that forecasts 2027. The reporting does not say what a new Congress would do with a new bill. Senate Commerce negotiations are technically ongoing, with key senators describing the state of talks as "not there yet" ^[4]. The split is on the record: Democrats want stronger guardrails around AI safety and transparency, while Republicans, led by Speaker Mike Johnson, worry that heavy-handed regulation could hamstring American companies in their race against Chinese competitors ^[9]. Johnson has been explicit that any regulatory approach must avoid undermining US competitiveness ^[13]. Trump dismissed existential-risk warnings from AI researchers as a "hoax" during September 2026 ^[10].