Consumer-facing AI use disclosure mandate

Movement organisations advocate for statutory or regulatory requirements that force AI systems to disclose their AI-ness to the human counterparty at the point of use — a chatbot must identify itself as a machine to the user; AI-generated content in a feed, an advertisement, or a political communication must carry a visible label; an algorithmic decision denying a benefit, a loan, a job, or a housing application must be identified as such to the affected person in plain language, along with the person's rights of contestation. The vehicle is a specific disclosure clause inside a broader AI or consumer-protection statute (EU AI Act Article 50, California SB 942, various US state deepfake and chatbot laws, sectoral rules under FTC and CFPB authority), operationalised through symbol/label standards developed with civil society and enforced through consumer-protection agencies.

An actor chooses this strategy because it operates at the individual-encounter surface where AI actually reaches the consumer — the moment when someone is deciding whether to trust a message, apply for a loan, or accept a decision — and it grants a specific enforceable right to know the consumer can act on. Disclosure creates the informational precondition on which every other consumer-side response depends: without knowing they are talking to a machine, a consumer cannot decide to talk to a human instead; without knowing an image is synthetic, they cannot discount it; without knowing an adverse decision was algorithmic, they cannot contest it under the procedural rights that attach to algorithmic decision-making in their jurisdiction. The strategy also converts an abstract worry about AI ubiquity into a concrete, testable, compliance-auditable requirement whose violation is actionable — a wedge that supports subsequent litigation and further regulation.

It trades directness for scope. Disclosure alone does not restrict what AI systems do — it only requires that they announce themselves. Consumers habituated to click-through disclosures reliably ignore them (privacy-notice fatigue is the paradigmatic case), and a well-crafted disclosure obligation can be satisfied by a footnote that reaches no one. The strategy is also implementation-heavy at the technical layer: reliable watermarking of AI-generated content is unsolved for most modalities, chatbot self-disclosure is trivial to evade, and cross-jurisdictional inconsistency in what must be labelled creates arbitrage. Without an enforcement machinery that treats disclosure as material — not a nominal disclaimer — the strategy stalls at compliance-theatre.

Ecology

Distinct from mandatory algorithmic impact assessment requirement. AIA is pre-deployment disclosure by the operator to a regulator: the operator produces documentation of its system's impact, files it with an oversight body, and the assessment becomes an artifact for regulatory review. This strategy is at-use disclosure to the consumer: the operator must inform the individual on the receiving end of the AI, at the point of encounter. The two register on different stakeholders (regulator vs. affected individual), operate at different points in the deployment cycle (before vs. during), and are typically packaged together in comprehensive AI laws — but their organising logics diverge and coalitions form around them differently.

Distinct from community-defined benchmarks and standards. Benchmarks measure AI performance; consumer disclosure marks AI presence. The two coexist: benchmarks tell the consumer how well an AI system performs on a task the community defined, disclosure tells the consumer that an AI system is what they are dealing with. Both are informational strategies against the opacity that AI deployment produces, but they inform on different axes.

A tactical prerequisite for organized consumer boycott of AI products. A consumer cannot organise around refusing AI products they do not know they are using. Mandatory disclosure creates the conditions on which a consumer-boycott strategy has a target to name: a service that consumers can identify as AI is a service they can refuse.

Fed by empirical audit and expose. Audits that document AI systems operating on consumers without disclosure — chatbots posing as human customer-service agents, adverse decisions made algorithmically without notice, synthetic content circulating unlabelled — provide the evidence base that moves legislators to enact disclosure requirements and regulators to enforce them.