Washington lawmakers approve HB 1170, a landmark bill requiring operators to disclose AI-generated or modified content.
What HB 1170 Requires
Washington State has passed HB 1170, a disclosure bill aimed at making the use of artificial intelligence in published content visible to the people who consume it. At its core, the law obligates operators to disclose when content has been generated or materially modified by AI systems. Rather than banning any particular use of the technology, it treats transparency as the mechanism for accountability: audiences are entitled to know whether what they are reading, watching, or hearing was produced or altered by a machine.
The distinction between "generated" and "modified" matters. A fully synthetic article and a human-written piece that has been rewritten, restructured, or enhanced by an AI tool can both fall within scope. Operators therefore need to think not only about content created from scratch by a model, but also about the everyday editing, summarizing, and augmentation steps where AI now sits quietly in the workflow.
Who Carries the Obligation
The bill places the duty on operators — the parties who publish or distribute the content — rather than on the underlying model providers. That framing is deliberate. The operator is the entity that decides what to release and in what context, so it is the natural point at which a disclosure can be attached and enforced. It also means the responsibility cannot be shifted onto a third-party tool: if you ship the output, you own the disclosure.
For any organization operating in or reaching audiences in Washington, this turns AI usage into a compliance question rather than a purely editorial one. Teams that have adopted AI informally, without tracking where and how it touches their output, will need a clearer internal picture before they can accurately disclose anything.
Preparing to Comply
Meeting a disclosure requirement is less about legal wordsmithing and more about knowing your own pipeline. Before you can label content honestly, you have to be able to answer where AI entered it. That argues for building disclosure into the production process rather than bolting it on at the end.
- Map where AI is used across drafting, editing, translation, image generation, and summarization.
- Define an internal threshold for what counts as "materially modified" so labeling is consistent.
- Decide on a clear, standard disclosure format and where it appears relative to the content.
- Keep records of AI involvement so a disclosure can be justified if questioned.
- Train contributors and reviewers so disclosure is a routine step, not an afterthought.
Why It Matters Beyond Washington
State-level rules rarely stay contained. Content published online crosses jurisdictions freely, and operators generally find it easier to apply one consistent disclosure practice than to maintain a patchwork keyed to each audience's location. A law like HB 1170 can therefore set a practical baseline that reaches well past the state that enacted it.
For the broader debate over AI accountability, the significance is in the approach. Disclosure does not judge whether AI-assisted content is good or bad; it restores the reader's ability to weigh the source and apply their own judgment. Building the habits to support that transparency now is the most durable way to stay ahead of similar requirements as they appear elsewhere.