Ethics Debate

last edited

The ethics debate concerns whether generative image models were trained on internet images without artists' consent. Disputes cover attribution, payment, employment and copyright. Courts in the US and UK, and regulators in the EU, now handle parts of the dispute. The open-source and specialized models and Midjourney are common targets of these claims.

Legal status in the US and UK

In the US, a lawsuit by artists against Stability AI is in discovery before a federal trial court. A discovery order was issued on 17 June 2026. Orders on discovery and sealing followed in March and August 2026.

In the UK, the court rejected a secondary infringement claim by Getty against Stability AI and left the legality of training unresolved. Getty obtained permission to appeal. The appeal opened on 11 November 2026. Getty argues a trained model can be an infringing copy.

EU rules

The EU AI Act requires general-purpose model providers to publish a summary of training content. The duty has applied since 2 August 2025 and covers models released under free and open-source licenses.

Enforcement became possible on 2 August 2026. Fines reach 3% of worldwide turnover or €15M, whichever is higher.

Grievances

Critics connect generation with lost freelance work, especially for designers and photographers. These concerns appear in artist departures from X, community sentiment about AI slop, and platform migration patterns.

Remedies and communities

Proposals include consent rules, dataset disclosure, bias audits, credit and payment. China has its own rules, covered in platform controversies and regulation there. Market effects are covered in AI-art freelance pricing and AI-created art market projections.

Community debate appears in core community conflicts and moderation problems. The aesthetic counterargument appears in visual philosophy.

Sources