Artificial Intelligence: a move towards EU legislation
On 9 December 2023, a long-awaited political agreement was reached between the European Parliament and the Council on the Artificial Intelligence Act. President von der Leyen said in a press release: “today’s agreement focuses regulation on identifiable risks, provides legal certainty and opens the way for innovation in trustworthy AI. By guaranteeing the safety and fundamental rights of people and businesses, the Act will support the human-centric, transparent and responsible development, deployment and take-up of AI in the EU”.
Details on the specifics of the Act remain reasonably scarce, however, the importance of what will become the world’s first comprehensive legal framework on artificial intelligence cannot be understated. The principal purpose of the Act is said to be to ensure that where AI is used in the EU, it is safe, transparent, traceable, non-discriminatory and environmentally friendly. At present, it is expected that:
• The structure of the Act will retain the same tiered, risk-based approach adopted in the original 2021 proposal.
• The Act will seek to distinguish AI from “simpler systems”, including by aligning with the Organisation for Economic Co-operation and Development (OECD) definition of AI, which is that: “an AI system is a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment”.
• There will be a “prohibited” category of AI which will ban, amongst other things, facial recognition systems in publicly accessible spaces (with certain exceptions), biometric categorisation systems using sensitive characteristics, emotion recognition in the workplace and in education, systems intended to manipulate human behaviour to circumvent free will, systems providing scoring based on social behaviour or personal characteristics, and AI used to exploit the vulnerabilities of people due to their age, disability, social or economic situation. There will likely be “carve-outs” for certain uses (e.g. for systems recognising when the driver of a car has fallen asleep), for certain types of law enforcement activity, and for defence purposes.
• There will be a second tier of controls imposing obligations which attach to high-risk AI systems due to their significant potential harm to health, safety, fundamental rights, environment, democracy and the rule of law. Controls will include mandatory fundamental rights impact assessments and are said to apply to, inter alia, the insurance and banking sectors. AI systems designed to influence the outcome of elections and voter behaviour are also classified as high-risk.
• EU citizens will have the right to make complaints about AI-based systems and to request explanations to support decisions affecting them which were made by AI systems.
• General-purpose AI systems which do not fall into the prohibited or high-risk categories must adhere to transparency requirements, which may include provision of technical documentation, compliance with copyright laws and disclosure of summaries of content used for training.
• The Act is said to promote and support innovation and SMEs by creating “regulatory sandboxes” and real-world testing initiatives to help develop and train innovative AI solutions before they hit the market.
• Stringent financial sanctions will apply for non-compliance with the act.
Before the Act becomes law, the agreed text must be formally adopted by both Parliament and Council. It therefore remains to be seen whether the AI Act is adopted in its present form. If it is, it will directly impact upon UK manufacturers selling into the EU, who will be required to fully understand, and comply with, the terms of the law. More widely, as the world’s first comprehensive legal framework on artificial intelligence the importance of the legislation cannot be understated: it will undoubtedly influence thinking around future legislation in the United Kingdom, North America, and beyond.