EU AI Act transparency obligations now in force, and disability and AI ethics expert warns UK businesses face extraterritorial compliance exposure regardless of Brexit

New rules under the EU AI Act took effect on 2 August 2026, requiring AI systems that interact with the public, including chatbots and AI-generated content, to clearly identify themselves as artificial. The obligations apply extraterritorially: any UK business whose AI systems, models, or outputs are used by or placed on the EU market falls within scope regardless of Brexit, and regardless of whether that business has any EU staff, servers, or legal entity. The EU AI Act is described in the source as the world's first comprehensive law regulating artificial intelligence. It classifies AI systems into four risk tiers. Eight practices judged to present unacceptable risk, including emotion-recognition technology in workplaces and classrooms and biometric systems inferring sensitive personal traits, have been banned outright since February 2025, with fines of up to €35 million or 7% of global turnover. Rules covering high-risk AI applications, including systems used in recruitment, worker management, and access to employment, were originally expected this August but have been pushed back to December 2027 following an EU-agreed delay. The transparency obligations now in force require chatbots to identify themselves as AI and AI-generated or altered images, audio, and video to be labelled as artificial. Rachael Mole, director of Moleworks Solutions Ltd and a disability inclusion consultant and AI ethicist, warns that UK employers using AI for recruitment screening already face obligations under the to make reasonable adjustments for disabled candidates, and that AI tools can penalise employment gaps caused by illness or disadvantage candidates who communicate differently. The has separately warned that AI recruitment tools can produce discriminatory job adverts and disadvantage candidates outside the system's training pattern.