The commission cannot use emergency rulemaking, so any rule it adopts would not apply to November’s election.
OKLAHOMA CITY — Oklahoma ethics officials are considering whether campaign ads should tell voters when artificial intelligence was used to create or substantially alter their content.
Any rule adopted by the Oklahoma Ethics Commission would come too late for the Nov. 3 general election.
Executive Director Lee Anne Bruce Boone has said the commission does not have emergency rulemaking authority. No draft rule was before commissioners Thursday, and they took no action.
Instead, staff outlined the legal questions Oklahoma would face in requiring AI labels on political advertising. The central issue is how to give voters useful information without interfering with political speech protected by the First Amendment.
Labels may be safer than bans
Staff said requiring a label generally stands a better chance in court than banning an AI-created campaign ad. A label allows the ad to run while telling voters how it was made.
But forcing campaigns to add certain wording also raises constitutional questions. A requirement could become legally vulnerable if it covers too much material or makes the label too large, lengthy or difficult to use.
The presentation, available through the commission’s meeting archive, raised several practical questions.
Should a label be required whenever AI is used, or only when it makes a meaningful change? Should basic editing be treated the same as a realistic fake video showing a candidate saying or doing something that never happened?
Oklahoma also would have to decide what the label should say and how it should appear on television, social media, printed material and other campaign advertising.
Staff said a simple, neutral statement that AI was used would raise fewer concerns than making a campaign describe its own ad as false or misleading. Any rule also would need to account for satire and parody.
Another question is whether Oklahoma could require greater transparency without making the Ethics Commission decide whether a political claim is true or false.
Current state rules generally require political ads to identify who authorized and paid for them. They do not require campaigns to disclose the use of AI.
Other states require labels
Commission staff reported in an August presentation that 31 states had laws addressing AI in political campaigns, up from six states in January 2024.
Disclosure was the most common approach, used in 28 states. The details vary, including what material must be labeled and whether violations bring civil or criminal penalties.
Staff also pointed to a federal court ruling that blocked a California law covering deceptive election material. The court found that the law restricted too much protected speech and raised concerns about its treatment of satire, parody and required labels.
The U.S. Supreme Court has not ruled specifically on required AI labels in political advertising.
A question for future elections
Under the commission’s rulemaking process, a proposed change would require a public hearing. A rule adopted by the commission would then go to the governor and Legislature when the 2027 legislative session begins.
It would take effect unless lawmakers rejected it through a joint resolution. The Legislature also could pass its own AI-disclosure law.
Until either happens, Oklahoma’s Ethics Rules do not require political ads to carry an AI label.
