Firms are dealing with a growing and increasingly complex set of AI rules. “New AI-specific requirements are being layered on top of existing, wider AI-relevant laws. In some cases, there is uncertainty about how and when some of these will apply,” says Flakoll.
The US has a patchwork of regulation, especially at state level. California, Texas and Colorado, for example, have AI laws scheduled to come into effect this year. “Colorado’s AI Act is one to watch because it deals with algorithmic discrimination and affects areas such as employment, lending and insurance, although its implementation has already been pushed back to June and could still shift,” she says. At the federal level, lawmakers have introduced a broad law that addresses many key AI policy issues such as harm to children and intellectual property protection.
President Trump has also recently issued a National Policy Framework for Artificial Intelligence, urging Congress to adopt a law that protects against certain potential AI harms while also stressing the importance of promoting innovation, including reiterating the president’s desire to pre-empt state laws that impose cumbersome requirements.
APAC continues to be active too. For example, South Korea and Vietnam have introduced new AI laws this year and China’s already sophisticated AI regulatory framework may further develop following a recent announcement that legislative research on AI law is a priority for 2026. Singapore and other countries also have more AI guidance in the pipeline.
In Europe, the focus is on the AI Act. A further key tranche of the Act’s requirements is currently scheduled to start applying from August 2026, including rules for high-risk AI systems. These can apply, for example, to AI systems used for evaluating creditworthiness, for life and health insurance risk assessment and pricing and certain recruitment and HR activities. However, it is unclear whether these obligations will enter into force as scheduled: the EU’s Digital Omnibus package of proposals includes a delay to applicability of the high-risk AI rules, but the proposals remain subject to the trilogue legislative process, which can be lengthy. “Many firms have decided to press on with the work needed to comply with high-risk AI rules and wider risk management expectations. We are helping clients work through these and, more widely, how to deal with this fast-moving landscape – especially for operations across borders – including what a holistic approach looks like,” Flakoll says.