KAIST

< AI x Failure Idea Competition Poster > 

KAIST announced on May 18th that the Center for Ambitious Failure (CAF) is holding the ‘2026 AI × Failure Idea Competition’ to explore potential failures and social risks that may arise in a future society where AI has become commonplace, through the eyes of citizens.

The theme of this competition is ‘2036, Why Did We Fail? Please write an incorrect-answer note from the future.’ Participants will assume a future in 2036 where AI is deeply integrated across society, and participate in a thought experiment to look back at the causes of possible failures and the signals we missed.

Notably, this year's competition features the introduction of the ‘Pre-Mortem (prior failure analysis)’ technique, which assumes failure in advance and analyzes its causes. By combining this method—traditionally used in the field of organizational psychology—with social imagination in the AI era, it has been expanded into a citizen-participatory platform to reflect on and prepare for future risks in advance.

The competition focuses on the fact that failures in the AI era are highly likely to occur through a complex intertwining of human judgment, social institutions, and changing relationships, moving beyond simple technological errors. Accordingly, it explains that the process of imagining and reflecting on future failures in advance can be a significant social learning experience in itself, rather than just simple technology forecasting.

This competition focuses on the insight to imagine failure and the capacity for logical backtracking, rather than specific technological ideas or implementation capabilities. Participants can submit a one-page proposal structured in three stages: ▲ Anticipated Failure, ▲ Diagnosis of Cause, and ▲ Countermeasures.

In last year's competition, 111 teams of undergraduate and graduate students from across the country participated, confirming a high level of interest in the issue of failure in the AI era and the coexistence of humans and technology. Consequently, the Center for Ambitious Failure plans to expand the target participants to the general public this year to explore future failures from the perspectives of a wider range of citizens.

Applications will be accepted online from May 18 to July 17. The top 10 teams selected through the first round of written evaluation will participate in the finals presentation and award ceremony to be held at the KAIST Daejeon Main Campus this coming August.

The total prize money is 10 million KRW, with the selection of 1 Grand Prize team (3 million KRW • KAIST President's Award), 1 First Prize team (2 million KRW), 2 Excellence Prize teams (1 million KRW each), and 6 Challenge Prize teams (500,000 KRW each).

Sung-Hee Cho, Director of the Center for Ambitious Failure, said, "This competition is not an event to simply predict the future, but a collective thought experiment to discover the signals we are missing right now by drawing out failures that have not yet arrived." He added, "I hope it serves as an opportunity for citizens who will live alongside AI to design the future themselves, rather than passively adapting to technological changes."

President Kwang Hyung Lee stated, "In the AI era, unexpected failures can emerge along with new possibilities." He added, "KAIST will continue to view failure not as a result to be hidden, but as an important learning process to prepare for the future, and will ponder better directions together with society."

Further details can be found on the KAIST Center for Ambitious Failure website (https://caf.kaist.ac.kr).