Appropriation encompasses a range of harmful extractive behaviours enabled or amplified by AI and automated systems:

• Creative appropriation. Companies harvest the writing, artwork, music, code, and other creative output of individuals and communities to train generative models, which then produce outputs that compete directly with the very creators whose work fed them. Artists and creators have raised concerns that generative AI systems are being trained using their works without their consent, credit, or compensation.

• Identity and likeness appropriation. AI tools are used to clone or replicate a person's voice, face, or persona without permission, enabling impersonation, fraud, and commercial exploitation. Courts in the United States and the European Union began classifying voice data as biometric property (not just creative output), thereby allowing individuals to claim ownership of their vocal signatures.

• Cultural appropriation. AI systems reproduce, commodify, or distort the knowledge, aesthetics, and cultural expressions of communities - particularly indigenous and marginalised groups - without their involvement, recognition, or benefit.

• Data appropriation. Personal data shared by users in one context is repurposed to train AI systems in another, without informed consent or awareness.

• Style appropriation. Even when specific works are not copied verbatim, AI systems learn and replicate individual artistic styles at scale. When users prompt AI tools with an artist's name hundreds of thousands of times, they leverage distinctive visual aesthetics that artist developed over decades of practice, enabling commercial-quality generation of style-mimicry without compensation or consent.

Why it matters

For the individuals and communities affected, their creative labour, cultural identity, and personal data represent livelihoods, rights, and deeply held expressions of self. When AI systems absorb and profit from these without consent, it:

• Undermines economic survival for writers, artists, musicians, journalists, and other creative professionals whose work is absorbed to build systems that compete with them.

• Erodes cultural sovereignty, particularly for indigenous communities whose knowledge and traditions are extracted without permission or benefit-sharing.

• Normalises exploitation at scale, treating the contributions of individuals as a commons to be harvested for corporate profit.

• Distorts power by concentrating the gains of AI development in the hands of technology companies while distributing none of the value to those whose contributions made it possible.

Impacts & consequences

When the ethical norms around appropriation break down, the consequences include:

• Economic harm. Creators find their market undercut by systems trained on their own work. The use of copyrighted works for AI training raises questions about whether AI-generated content displaces the market for originals, threatening the futures of creative professions.

• Reputational and psychological harm. Victims of voice or likeness appropriation face fraudulent impersonation, deepfake scams, and loss of control over their own public identity. Steve Harvey described scams using his likeness as at "an all-time high," expressing concern for fans defrauded by AI-generated impersonations of him.

• Cultural harm. Communities see their heritage commodified or misrepresented, stripped of context or sacredness.

• Erosion of trust. Both in AI systems and in the digital platforms where people share creative and personal content.

• Legal and regulatory uncertainty. The US Copyright Office released findings in 2025 concluding that using copyrighted materials for AI model development may constitute prima facie infringement, warning that "transformative use" arguments are not inherently valid, yet enforcement remains inconsistent and inaccessible to most individual victims.

• Chilling effect on creativity. Artists may withhold or obscure their work to protect it from being ingested, impoverishing the broader cultural commons.

Causes & contributing factors

• The absence of opt-in consent frameworks. Most AI companies currently operate under implied consent, essentially arguing that posting art online constitutes permission for any use. Artists and creators increasingly reject this interpretation.

• Legal lag. Intellectual property law was designed for human-to-human copying and does not clearly address AI-scale ingestion and style replication.

• Commercial incentives. AI model performance improves with more training data, creating powerful financial incentives to maximise data collection regardless of provenance.

• Opacity. Training datasets are rarely disclosed, making it impossible for creators to know whether or how their work has been used.

• Platform economies. The structure of social media and creative platforms encourages broad sharing, exposing creative work to scraping at scale.

• Weak enforcement. Even where laws exist, litigation is expensive and slow, and the burden falls on individual victims.

• Cultural insensitivity. AI developers frequently lack awareness of the significance or sovereignty of the cultural materials they ingest.