AI Search and the Right to Access Information
AI search systems that provide direct answers rather than lists of links change how information is accessed and what perspectives are presented. As AI replaces traditional search for many users, the speech implications of how AI answers questions become significant.
Search engines have always shaped what information users can find. AI-powered search represents a qualitative change: instead of providing ranked links, AI search provides synthesized answers — incorporating its own judgments about what is accurate, relevant, and appropriate to present.
When AI search provides an answer rather than links, several free speech concerns arise:
Synthesis effects: AI systems that synthesize information from multiple sources inevitably emphasize some perspectives over others in ways that are not transparent to users.
Accuracy and bias: AI systems that refuse to return results for certain queries, or that systematically return results reflecting particular viewpoints, effectively act as information gatekeepers — but with less transparency than editorial decisions made by human journalists or librarians.
De-indexing and suppression: Search engines and AI answer systems can effectively remove information from accessible public discourse by declining to index or return it. This de-facto suppression is not constrained by the First Amendment for private companies, but has significant implications for information access.
Concentration effects: As AI search is dominated by a few large companies, the idiosyncratic content policies of those companies have outsized effects on global information access.
The right to receive information — the flip side of the right to speak — has been recognized by courts as a First Amendment value. Whether the concentration of information access in AI systems that make opaque, unaccountable choices about what to present raises First Amendment concerns for government systems is a developing question.