Microsoft Is Offering to Help U.S. Politicians Crack Down on Deepfakes

(Bloomberg) -- Microsoft Corp. is offering US politicians and campaign groups a tool aimed at helping fight the rise of deepfakes ahead of the 2024 presidential election.

The new tool will allow parties to authenticate their videos and photos with so-called watermark credentials, Microsoft President Brad Smith said on the company’s blog early Wednesday. Using cryptography, the service would create a permanent record of an image or video, allowing anyone on the web to see whether it has been digitally altered or generated using artificial intelligence over time, Smith said.

Initially, Microsoft plans to offer the tool for free to political candidates and then may eventually extend it to more groups after the 2024 US elections, a person familiar with the situation said, asking not to be identified because the information isn’t public.

Microsoft’s offer comes as concerns mount over AI-generated images and videos and their potential to supercharge misinformation as US citizens prepare to vote a new president into office in November 2024. Several political deepfakes have already emerged: Ron DeSantis was chastised in June for publishing fake images on the social media platform X of Donald Trump cozying up to Dr. Anthony Fauci — an attempt to mock his Republican party challenger. In April, the Republican National Committee circulated a fake video depicting an apocalyptic US under President Joe Biden’s rein.

Related:ChatGPT Confident in Its Software Development Capabilities, Survey Finds

The extent to which politicians and their campaign groups will actually take up Microsoft’s offer and use its tool remains unclear. A deepfake offering has become an election tradition of sorts for Microsoft, which in 2020 announced a tool for detecting them and similarly offered politicians a means of authenticating their content.

And while Microsoft is warning of deepfakes and offering to help fight them, the company is also heavily investing in the space of generative AI. Microsoft has pumped $10 billion into OpenAI, the creator of one of the most popular AI image generators, Dall-E.

Just weeks before the Biden video was released, a manipulated clip went viral of Senator Elizabeth Warren saying Republicans should be banned from voting.

A group of US senators introduced a bipartisan bill in September that would ban the use of deepfakes to influence elections or solicit political funding.

Microsoft also posted a new threat intelligence report that doubled down on its previous concerns about foreign actors’ attempting to influence campaigns. It’s likely that China, Iran, and Russia will try to meddle with US politics in the next year, the company said in the report. 

Related:Will You Be Replaced By Your AI Clone? Maybe One Day, But Not Quite Yet

The 2022 US midterm elections were widely expected to be plagued with cyberattack attempts and influence operations out of Russia, Microsoft said in the report, but no major incidents were reported. Still, Russia remains “the most committed and capable threat to the 2024 election,” the company said, adding that the country would likely use synthetic content to sow discord on social media.

“The Kremlin likely sees next year’s contest a must-win political warfare battle determining the trajectory of support to Kyiv and the outcome of the Ukraine War,” Microsoft said in its report.