Regime divergence, not any single AI statute, is now the binding compliance constraint.
The EU imposes strict product liability on all software, the UK excludes standalone software and regulates embedded AI through sectoral product-safety regulators, South Korea relies on effort-based self-regulation, and US states run their own patchwork — forcing multinationals to maintain a distinct compliance posture for every jurisdiction.
The same conclusion keeps arriving from across the workspace's research — 1 topics independently instantiate this theme. Filter the evidence by where it came from:
Korea's effort-based 'high-impact' model demands its own posture apart from the EU's 'high-risk' duties.
Simultaneous EU, UK, and US regimes built on different philosophies make maintaining a distinct posture per jurisdiction — not any one statute — the binding constraint.
The UK consultations codify a product-safety regime excluding standalone software while the EU's PLD sweeps it in — regime divergence, not any single statute, dictates the compliance posture.
Even a maximum-harmonization directive is transposing into divergent national AI-liability regimes ('Same AI product, three different liability regimes'), extending the jurisdiction-by-jurisdiction compliance constraint into product liability.