How to Trick Users into Filling in Your Boring AI Ethics Survey Using React

We’re at a turning point where it’s easier to claim that real images are fake than to prove that images are generated by AI. When it’s increasingly easy to generate convincing depictions of individuals doing or saying things that they never did we’re stuck in a situation where we can no longer agree on a shared reality. What does this mean for using audio as evidence in court if anyone can mimic speech? How can we make informed voting decisions when digital tools can make politicians say or do anything? To explore these issues, I designed an experiment called the "Turing Twest (Tweet + Test)," which challenges users to differentiate between tweets from well-known public figures (like Elon Musk, Donald Trump, and Taylor Swift) and AI-generated tweets. The design of the test is made to mimic Twitter’s UI using React and AntD. The game-like interface as opposed to a survey is there to motivate users to respond and replay multiple times. The results showed that people initially struggled to identify real tweets from fake ones, but also revealed an interesting detail on how we can improve media literacy in the attention economy.