Synthetic everything is warping America

A shadowy polling firm, Median Strategies, confessed to faking survey results in Wisconsin and Los Angeles races, with 21-year-old Rahil Prakash revealing himself as the operator in an interview with The Guardian. The incident highlights how AI-enabled synthetic content is warping political, business, and cultural signals, as seen in deepfake robocalls, fake campaign ads, and a $25 million fraud against Arup.

A shadowy polling firm confessed this week to faking survey results in two marquee races — then claimed the entire operation was a "short-term social experiment" on misinformation. Why it matters: An AI-enabled flood of synthetic and outright fabricated content is wreaking havoc on reality, poisoning the signals people rely on to understand politics, business and culture. Catch up quick: Median Strategies, a previously unknown polling outfit with little public footprint, surfaced this summer with two blockbuster surveys that were quickly absorbed into the political conversation. In Wisconsin's Democratic gubernatorial primary, Median's fake poll showed democratic socialist Francesca Hong leading Milwaukee County Executive David Crowley by 23 points — reinforcing a broader consensus that cast Hong as the clear favorite. Crowley ultimately won by half of a percentage point. In Los Angeles, Median reported Mayor Karen Bass leading city council member Nithya Raman by nearly 12 points. Bass cited the poll as evidence her campaign was "gaining momentum," only to delete her post after the survey was exposed as bogus. 21-year-old recent college grad Rahil Prakash revealed himself as the person behind Median's operation in an interview with The Guardian. The intrigue: The rise of prediction markets has added a financial incentive for political manipulation. The big picture: America's fractured information landscape was already ripe for political manipulation. The cost of deception is now plummeting thanks to AI, allowing bad actors to manufacture convincing voices, images and identities at enormous scale. Robocalls using AI to imitate Joe Biden's voice discouraged thousands of people from voting ahead of the 2024 New Hampshire Democratic primary. Political groups have begun weaponizing AI deepfakes in campaign ads, exploiting a patchwork of regulation to depict opponents in compromising situations that never happened. Businesses are also grappling with terrifying new risks thanks to the mass production of fake people and fake content. North Korean operatives have built "a secret workforce" inside U.S. companies, using stolen identities, AI tools and U.S.-based facilitators to land remote jobs, according to a Wall Street Journal investigation. Global consultancy Arup lost more than $25 million after an employee joined a video call with deepfake versions of company executives and followed their instructions to transfer money to various bank accounts. Zoom out: The same synthetic signals are increasingly warping everyday life, manufacturing the appearance of popularity, authority and consensus. It's playing out on an internet where a majority of traffic now comes from bots. Companies are seeding the web with product rankings that favor their own offerings, knowing chatbots may absorb those rankings and repeat them to consumers. The Atlantic dubbed the practice "Sloptimization." Marketers are paying networks of accounts to manufacture online buzz around songs, movies and celebrities, Vulture reported — generating enough artificial momentum to trigger algorithms and create genuine attention. Not even the weather has been spared: Alabama meteorologist James Spann said he fielded more than a dozen questions about an imaginary late-August "cold wave" this week after AI engagement farms churned out bogus forecasts online. The bottom line: The public is already struggling to navigate the first wave of synthetic reality. As AI accelerates, authenticity and trust will only grow more precious.