Google is soon enforcing a rule that political ads on its platforms must indicate if photos and audio have been generated by using AI.
In response to the increasing presence of technologies generating synthetic content, a Google spokesperson told the BBC that regulations had been instated.
The transition is slated for November, roughly twelve months prior to the subsequent US presidential election.
There is concern that AI will significantly amplify false information about the election.
Google's current rules forbid altering digital media in ways meant to deceive or misinform people concerning political, social, or public matters.
This upgrade necessitates that ads pertinent to elections to be "clearly disclosed" when using "synthetic content" that appears as if it is replicating real people or situations.
Google proposed the use of tags such as "this image does not show actual events" or "this video content was invented artificially" as indicators.
Google has prohibited any demonstrably false claims that could hurt trust in the election process, according to its ad policy.
Google mandates that political advertisements must show who paid for them, and the ads library holds all related data about them.
Disclosures of digitally altered content in election ads should be highly visible, and located in places where it is probable to be seen.
Labels would be required for synthetic imagery and audio depicting a person doing or saying something they did not, as well as for false depictions of events.
In March, an AI-generated fake image depicting former US President Donald Trump's arrest circulated on social media.
In March, a deepfake video circulated in which Ukrainian President Volodymyr Zelensky appeared to be speaking of yielding to Russia.
This video is unable to be played.
James Clayton of the BBC tries out a deepfake video detector.
In June, a campaign video released by Ron DeSantis showed images that had indications of having been generated by Artificial Intelligence (AI).
A tweet shared a video featuring what seemed to be edited images of President Trump embracing Anthony Fauci, a major player in the United States' COVID-19 response team, with kisses on both cheeks.
Experts in AI have informed the BBC that though fake visuals are not novel, the speed of progress in the field of generative AI and the opportunities for misuse is something to worry about.
Google declared that it remains devoted to investing in technology for the discovery and eradication of such content.
Comments