Campaigns that use artificial intelligence in an attempt to sway voters would soon be required to reveal that fact, under a bipartisan bill Oregon lawmakers are considering this year.
Senate Bill 1571 would require campaign materials — from physical fliers to online videos — to disclose the use of any artificial intelligence used to depict a person’s voice or image. It’s one of dozens of similar proposals that have emerged around the country, as state and federal officials work to tamp down use of emerging “deep fake” technology ahead of this year’s elections.
“It’s important that the state of Oregon keeps up with the times”, said state Sen. Aaron Woods, D-Wilsonville, a retired Xerox executive who introduced the bill. “The bill will build awareness.”
Oregon hasn’t seen high-profile instances of artificial intelligence in political communications yet. But as technology quickly improves, AI has become a factor on the national stage — including in robocalls using the faked voice of President Joe Biden in New Hampshire and a television ad using AI to mimic the voice of former President Donald Trump. The Federal Communications Commission outlawed the use of AI in robocalls last week.
Woods’ bill, which was scheduled to receive a public hearing Tuesday afternoon, has support from Democrat and Republican lawmakers. Under an amendment Woods expects to move forward, it defines “synthetic media” as an image, audio recording or video of a person that “has been intentionally manipulated with the use of artificial intelligence techniques or similar digital technology” and which gives voters a false impression of events.
Campaigns using that material would be required to disclose it. Violations could result in a lawsuit from the Oregon Secretary of State and a maximum $10,000 fine. The bill includes exemptions for media organizations that report on campaign ads that feature AI.
The bill would take effect immediately, putting its provisions into place well ahead of Oregon’s May primary election.
The regulations in the bill are not as detailed or strict as those in some other states. A law in Michigan, for instance, requires that a message about use of AI be written in the same sized font as the majority of texts when used in print ads, and be put on the screen for at least four seconds in the case of video ads.
Public Citizen, a national advocacy group pushing AI disclosure rules around the country, advocates stronger regulations, including that AI disclosures be written in the same-sized font as the largest writing in print communications, and be displayed the duration of a video.
More than three dozen states have introduced or passed bills to ban so-called “deep fake” videos in politics or require disclosure when they are used, according to the organization. Those efforts have been broadly bipartisan.
It’s not clear from HB 1571 exactly how Oregon campaigns will be required to reveal use of AI if the bill passes, other than that they must state that the material has “been manipulated.” Woods said he had not considered specific regulations around how disclosures must be presented, saying those details could be hammered out in a future legislative session. But he characterized his proposal as popular among colleagues, no matter their party affiliation.
“It’s bipartisan and bicameral, so if you are against this… I certainly want to hear why,” he said in an interview. “But no, we’ve not had any opposition or negativity.”
Among supporters is Secretary of State LaVonne Griffin-Valade, the state’s top election official, a spokesperson said Tuesday.
As of early afternoon, written testimony filed on the proposal was supportive or neutral, with “good government” groups like Common Cause Oregon and the League of Women Voters of Oregon voicing their support.
