Deepfake it till you make it

Prime Minister Narendra Modi greets people during the Hindustan Times Leadership Summit, in New Delhi, on Saturday, Nov. 4, 2023.

Prime Minister Narendra Modi greets people during the Hindustan Times Leadership Summit, in New Delhi, on Saturday, Nov. 4, 2023.

ANI via Reuters Connect
AI-generated songs featuring the (fake) voice of Indian Prime Minister Narendra Modi are taking over Instagram. One video, a would-be Modi cover of a popular Bollywood song, was viewed 3.4 million times on Instagram Reels, India’s leading social video platform after it banned TikTok in 2020. The online magazine Rest of World notes that these songs, translated into India’s many regional languages, could break down a language barrier for the Hindi-speaking Modi ahead of the 2024 general election.

While deepfake technology is typically associated with deceit – tricking voters and disrupting democracy – this seems like a more innocuous way of influencing global politics. But the Indian government hasn’t embraced this technology: It recently considered drastic action to compel WhatsApp parent company Meta to break the app’s encryption and identify the creators of deepfake videos of politicians. Deepfakes could, in other words, have a tangible impact on the world’s largest democracy.

In the United States, it’s not just politicians who have clashed with AI over this brand of imitation, but celebrities too. According to a report in “Variety,” actress Scarlett Johansson has taken “legal action” against the app Lisa AI, which used a deepfake version of her image and voice in a 22-second ad posted on X. The law may be on Johansson’s side: California has a right of publicity law prohibiting someone’s name, image, or likeness from being used in an advertisement without permission. In their ongoing strike, Hollywood actors have also been bargaining over how the studios can, if at all, use their image rights with regard to AI.

Deepfake technology is only improving, making it ever more difficult to determine when a politician or celebrity is appearing before your eyes – and when it’s just a dupe. In his recent executive order on AI, President Joe Biden called for new standards for watermarking AI-generated media so people know what’s real and what’s computer generated.

That approach – akin to US consumer protections for advertising – has obvious appeal, but it might not be technically foolproof, experts say. What’s more likely is that the US court system will try to apply existing statutes to new technology, only to reveal (possibly glaring) gaps in the laws.

Generative AI and deepfakes have already crept into the 2024 election, including a Republican National Committee ad depicting a dystopian second Biden term. But look closely at the top-left corner of the ad toward the end, and you’ll notice the following disclosure: “Built entirely with AI imagery.” Surely, this won’t be the last we see of AI in this election – we’ll be keeping an eye out for all the ways it rears its head on the campaign trail.

More from GZERO Media

Aubrey and Taylor Endicott cast their votes while their children, from left, Sterling, 5, Adelaide, 3, and Lincoln, 7, wait patiently under the voting tables at Shawnee Heights Middle School on Tuesday, Nov. 5, 2024.
USA Today Network via Reuters

For the US and other countries with elected leaders, Tuesday’s election represents a referendum on the future of democracy. Will it come away battered, or will it remain intact?

The Meta logo is seen on a mobile phone with the Chinese flag in the background in this photo illustration.
Photo by Jaap Arriens / SIpa USA via Reuters

But because of Meta’s openness, Chinese researchers were able to develop their own AI model — for military use — using one of Meta’s Llama models.

An FPV drone with an attached portable grenade launcher is seen during a test flight conducted by Ukrainian servicemen of the 'Bulava' Unmanned Aerial Vehicles Unit of the Separate Presidential Brigade at their position near a frontline, amid Russia's attack on Ukraine, in Zaporizhzhia region, Ukraine, on Oct. 11, 2024.

REUTERS/Stringer/File Photo

Ukraine is reportedly using new AI-powered drones to fly explosives toward Russian targets.

Pens for the diabetes drug Ozempic sit on a production line to be packaged at the Danish drugmaker Novo Nordisk's site in Hillerod, Denmark.
REUTERS/Tom Little

Armed with a bunch of Ozempic money, the Novo Nordisk Foundation — along with Denmark’s Export and Investment fund — is bankrolling a new AI supercomputer called Gefion, which launched on Oct. 23 and is run by a new company called the Danish Centre for AI Innovation.

DALL-E

While neither Vice President Kamala Harris nor former President Donald Trump has given much attention to artificial intelligence on the campaign trail — and AI hasn’t completely disrupted the election process as some experts feared — there are still important questions surrounding AI and the election.

US Capitol building at in the morning sun. Washington DC, USA The US Capitol building in the early morning at sunrise.

While eyes around the globe will be on the US presidency this Election Day, there are consequential races further down the ballot that will determine how much power Kamala Harris or Donald Trump will wield.