In 2018, Cambridge researchers Jon Roozenbeek and Sander van der Linden did something counterintuitive with the fight against fake news: they built a game that makes you the villain. Bad News hands you a troll account and walks you through building a disinformation empire — impersonating officials, stoking outrage, deploying fake experts. Fifteen minutes, free, in a browser. More than a million people have played it.
The theory behind it is older than the internet. William McGuire proposed attitudinal inoculation in the 1960s: expose someone to a weakened form of an attack on their beliefs, and they build resistance to the strong form, the way a vaccine uses a weakened pathogen. Bad News weaponized the idea in reverse — the weakened dose is performing the manipulation yourself, in a sandbox where the only victim is fictional.
A 2024–25 replication with Indian participants found no measurable improvement in telling true from false headlines. Effect sizes in the successful studies are modest, and they decay over weeks without booster sessions. The honest summary of the literature: inoculation works, somewhat, for a while, when the training matches the audience's actual media world. Generic doses fade. Google's Jigsaw unit learned the same lesson in the field — their prebunking ads measurably worked in Poland and did close to nothing in Slovakia, where the videos ran dubbed instead of localized.
We read those caveats as instructions rather than bad news. Match the scenarios to the person's real life. Repeat the reps. Keep updating the tactic file. That's the design brief for the 30-Day Path, and it's why our simulations run the boss, the mother, and the dealership rather than headlines — nobody's headline guilt-trips you at Sunday dinner.
If you want the mechanism firsthand, play Bad News first — it's free and it isn't ours. Then take our three-minute test and notice which strings still get past you.