Skip to content
One of the quietest but most consequential AI ethics questions is also one of the oldest: who gave permission for the data? In March 2026 two governments gave very different answers.
The UK holds the line
After more than eleven thousand consultation responses, many of them from writers, musicians and artists rejecting a broad text-and-data-mining opt-out, the UK government published its report on copyright and AI and largely kept the status quo. No new regulator, and for now no statutory transparency duty on AI developers to say what they trained on.
The EU leans the other way
The EU’s approach points toward disclosure: obligations on general-purpose AI providers to summarise training data and respect rights reservations. UNESCO, reporting the same month on how its AI ethics recommendation is being implemented across more than a hundred countries, found awareness rising but action uneven.
Why consent is an ethics issue
Copyright is the legal frame, but the ethical core is consent and provenance. People whose work, words or images trained a model have a reasonable interest in knowing that, and in some say over it. Organisations buying or building AI inherit that question. If you cannot say where a model’s training data came from, you cannot fully answer for what it produces.
Good data governance is where this starts: knowing what data you hold, where it came from and what you are entitled to do with it. See how we help organisations use AI they can account for.
Sources: UK report on Copyright and AI (Hogan Lovells) · UNESCO Ethics of AI
