One in five members of Congress holds household investments in AI infrastructure or model developers. This is not a scandal — it is a structural explanation for why the US has no meaningful AI regulation, and why the EU's AI Act remains the only comprehensive framework. For UK mid-market legal, insurance, financial services and accountancy firms with US operations, this matters urgently. You face a two-tier regulatory environment: the EU AI Act creates hard obligations; US federal rules remain vague and fragmented. That asymmetry creates real compliance costs. Your US-facing AI tools must meet EU standards that your American competitors can ignore. Until Congress acts — which financial incentives actively discourage — you will operate in this friction zone.

This story reveals something the industry has been reluctant to admit: AI regulation is stalled not because of technical complexity, but because powerful people profit from the delay. Harvey, Legora and Luminance have all raised significant capital betting on a US regulatory environment that remains permissive. Venture money flows fastest where oversight is lightest. Meanwhile, the FCA's approach to AI governance (embedded in Consumer Duty PS22/9 and emerging PRA SS1/23 guidance) assumes you are making deliberate, documented choices about AI risk. That assumption breaks down when your American equivalents face no equivalent pressure. The market is optimising for legal uncertainty, not legal compliance.

Trovix's position on this is clear: AI governance cannot be outsourced to the vendors. Harvey's promise — that large language models can reliably review contracts or predict litigation risk — depends on regulation staying loose enough that 'reliably' never gets defined. Luminance's appeal rests on opacity that regulators have not yet forced into the open. The honest alternative is to treat AI as a tool that augments human judgment under documented governance, not as a replacement for it. That means Trovix Watch is not optional — you need live monitoring of regulatory change across jurisdictions because Congress's delays create gaps you must fill. It means Trovix Audit becomes your defence: a compliance dashboard that proves you implemented AI with proportionate oversight, even while the baseline rules remain contested.

What you should do now: first, map your AI tools against EU AI Act classifications, not US law. The EU standard is your regulatory floor. Second, document your implementation decisions — why you chose that vendor, what safeguards you added, how you tested output quality. The FCA and SRA will ask. Third, assume Congress will eventually regulate — probably in a way that makes current practice look reckless. Build accordingly. Mid-market firms that treat AI as a speed play often learn too late that speed created liability. The firms that will win the next five years are those building AI governance now, while competitors still assume American permissiveness will last forever.

Source: Washington Post