AI-Generated Disinformation: The Crisis That Wasn't

Despite warnings from experts, AI-generated deepfakes have not yet created the massive truth catastrophe many feared. Amid the intense news cycles of recent times, media outlets and tech platforms have improved at spotting and debunking AI misinformation quickly.

Virality of Misinformation

Most viral falsehoods during breaking news events are not from doctored media but from manipulated contexts, such as misleading captions. An example is the aftermath of the Trump assassination attempt, where a doctored image showing Secret Service agents smiling while lifting Trump to his feet was quickly debunked by major news outlets and fact-checking services.

Rapid Fact-Checking Response

Mainstream news outlets like Reuters, The AP, Politico, BBC, and CNN, along with fact-checking sites such as Factcheck.org, Verify, and Politifact, used standard methods like reverse image searches to debunk the fake image rapidly. Despite these efforts, doctored photos and videos can still spread if tech platforms aren't swift in policing them.

The Role of Tech Platforms

The fake image of the smiling Secret Service agents still exists on X (formerly Twitter) but now has a Community Note attached. While these notes can be helpful, they are not always added quickly or widely enough to prevent the spread of misinformation. X has become a breeding ground for misinformation, often picked up and spread on smaller platforms where it goes viral.

Persistent Misinformation

A deepfake PBS News video of Biden cursing continues to circulate on X, but PBS quickly pointed out its inauthenticity. Broader conspiracies, often unrelated to doctored media, have also proven challenging to stop. Following the Trump assassination attempt, numerous conspiracies about the gunman's identity and motives gained traction online, independent of any doctored photos.

Challenges with AI Chatbots

While news outlets and fact-checkers have become adept at debunking AI-generated deepfakes in real-time, customized responses from AI chatbots about breaking news are much harder to police, posing a new challenge in the fight against misinformation.