Forget the infinite scroll of “what will deepfakes do in 2024?” thinkpieces and social media posts. 

What are deepfakes doing to the world right now?

Over the weekend, I read three critical and timely stories providing a snapshot of the material impacts deepfakes and misused generative AI are having on our world today. These articles, each focused on a different (mis)use of AI, reveal the unsettling reality that this technology is actively eroding truth and trust across many facets of society.  

They are proof positive that we cannot afford to underestimate or ignore the dangers already unfolding in front of us, as they can happen anywhere and to anyone.

Advertisers Left on X Push Crypto Scams, AI ‘Undressing’ Apps, users say after Musk's outburst
(Business Insider)

Creating non-consensual deepfake pornography of anyone in 2024 is as easy as pressing a button. Ads promising the “AI undressing” of anyone have littered adult entertainment sites for the last several months. Now, after nearly every major advertiser has fled X (formerly known as Twitter), the site is relying partially on these ads to stay afloat. These ads aren’t going away on X or anywhere for the time being, and they’re increasing the access and awareness of such tools and tech to a much wider audience.

The ramifications of these tools’ ubiquity are already apparent. Kids are being targeted by their peers in school. Authorities are late to act in charging perpetrators, or not acting at all. Victims have no way to fight back on a state or federal level, though a legal fix for this is currently being mulled in Congress.

For now, the hundreds of millions of users on one of the most popular social media platforms in the world are bombarded with the opportunity to undress another person without their consent, all with the use of AI. Should they choose to do so, they’re unlikely to be punished while the victim deals with a gross invasion of their privacy and well-being. 

Main accused in Rashmika Mandanna deepfake video case arrested, says police
(The Indian Express)

India, on the other hand, has stepped up their enforcement of non-consensual deepfakes, including a recently viral one of actress Rashmika Mandanna. Based on our findings, the country has so far implemented (or at least promised to implement) the strictest enforcement and penalties against those using deepfakes for the intent to cause harm or disinformation. Such actions come after many similar notable deepfake-related incidents, including a few of Prime Minister Narendra Modi. Other countries are struggling to follow suit.

Winner of Japan’s Top Literary Prize Admits She Used ChatGPT
(Vice) 

Though author Rie Kudan wrote only five percent of her Akutagawa Prize-winning novel, Tokyo Sympathy Tower, using an LLM, this still marks the most significant (and quasi-deceptive) usage of generative AI in a major artistic work. The Akutagawa Prize is one of the most prestigious literary prizes in Japan, akin to the National Book Award.

The win and scandal also come at a time when authors around the world are railing against the usage of their works to train AI models, finding AI rewrites of their books on Amazon, and being told they should offer their books for training by default. The prizewinning author is defending her usage of LLMs in the writing process, especially considering the book itself deals with themes related to AI.

The likelihood of authors around the world currently using these tools to augment or outright replace entire sections of their literary works is more or less a given. It’s only a matter of time before AI-generated text makes it to the award stages here in the West. How judges and the greater literary world will react — especially at a time when their professions are under threat from tools used by millions — will determine the next steps of regulation on this specific front.