Years before the release of ChatGPT, my research group, the University of Cambridge’s Social Decision Lab, wondered whether neural networks might generate false information. To achieve this goal, we trained ChatGPT’s predecessor, GPT-2, on examples of popular conspiracy theories and then asked it to generate fake news for us. He gave us thousands of misleading but acceptable news. A few examples: “Some vaccines are full of dangerous chemicals and toxins” and “Government officials have manipulated stock prices to hide scandals.” The question was, does anyone believe these claims?
We developed the first psychometric instrument to test this hypothesis, which we called the Misinformation Sensitivity Test (MIST). In collaboration with YouGov, we used AI-generated headlines to test how susceptible Americans are to AI-generated fake news. The results were alarming: 41 percent of Americans mistakenly thought the vaccine title was correct, and 46 percent thought the government was manipulating the stock market. Another recent study published in the journal sciencenot only showed that GPT-3 produced more convincing false information than humans, but also that people could not reliably distinguish between false information produced by humans and artificial intelligence.
My prediction for 2024 is that AI-generated disinformation is coming to an election near you and you probably won’t even know it. In fact, you may have already been exposed to a few examples. In May 2023, a viral fake story about a bombing at the Pentagon was accompanied by an AI-generated image showing a large cloud of smoke. This caused a public protest and even a drop in the stock market. US Republican presidential candidate Ron DeSantis used fake images of Donald Trump hugging Anthony Fauci as part of his political campaign. By combining real and AI-generated images, politicians can blur the lines between reality and fantasy and use AI to bolster their political attacks.
Before the explosion of artificial intelligence, cyber advertising companies around the world needed to write their own misleading messages and employ human troll factories to target people at scale. With the help of artificial intelligence, the process of generating misleading news headlines can be automated and armed with minimal human intervention. For example, microtargeting—the practice of targeting people with messages based on digital tracking data, such as their Facebook likes—was already a concern in past elections, despite the major hurdle of needing to generate hundreds of different The message was similar. To see what works on a certain group of people. What was once labor intensive and expensive is now cheap and readily available with no barriers to entry. Artificial intelligence has effectively democratized the creation of disinformation: anyone with access to a chatbot can now model a specific topic, whether it’s immigration, gun control, climate change, or LGBTQ+ issues. and generate dozens of very convincing fake news stories in minutes. . In fact, hundreds of AI-generated news sites are now popping up and publishing false stories and videos.
To test the effect of such artificial intelligence-generated misinformation on people’s political preferences, researchers at the University of Amsterdam created a fake video of a politician insulting his religious voter base. For example, in the video, the politician joked, “As Christ would say, don’t crucify me for that.” The researchers found that religious Christian voters who watched the fake Dip video had more negative attitudes toward politicians than did those in the control group.
It’s one thing to mislead people with AI-generated false information in experiments. Another experiment with our democracy. In 2024, we will see more deepfakes, voice simulations, identity manipulation, and artificial intelligence-generated fake news. Governments will severely limit – if not ban – the use of artificial intelligence in political campaigns. Because if they don’t, AI will undermine democratic elections.
