"2036, Why Did We Fail? Please Write an Answer Sheet from the Future."
KAIST announced Tuesday that its Failure Lab will host the "2026 AI×Failure Idea Contest" to explore potential failures and social risks in a future society where AI has become commonplace, viewed through the eyes of citizens.
Participants will engage in a thought experiment, imagining the year 2036 when AI has become entrenched across society, and tracing back the causes of possible failures and the signals we may have missed.
The contest focuses on the premise that failures in the AI era are likely to extend beyond simple technical errors, arising from a complex interplay of human judgment, social institutions, and changing relationships.
Rather than emphasizing specific technical ideas or implementation capabilities, the contest centers on the insight to imagine failure and the ability to logically trace its origins. Participants are asked to submit a one-page proposal structured around three stages: a foreseen failure, a diagnosis of its causes, and proposed countermeasures.
This year, the Failure Lab has expanded eligibility to all citizens, aiming to explore future failures from a wider range of perspectives.
Applications will be accepted online from May 18 to July 17. The top 10 teams selected through the first-round written review will participate in the final presentation and awards ceremony at KAIST's main campus in Daejeon in August.
The total prize pool is 10 million won. One grand prize team will receive 3 million won and the KAIST President's Award, one team will receive the top excellence award of 2 million won, two teams will receive excellence awards of 1 million won each, and six teams will receive challenge awards of 500,000 won each.
"This collective thought experiment to discover the signals we are currently missing will serve as an opportunity for citizens living alongside AI to actively respond to technological change," said Cho Sung-ho, director of the KAIST Failure Lab.
