Cultural Alignment

AI safety concept

Reward Hacking

Reward hacking occurs when an AI exploits or changes how success is measured to score well without achieving the intended goal.

Scenario file

Scenes in this index

18

Sort scenes

  1. Still from South Park illustrating The List

    South Park2007

    The List

  2. Still from Futurama illustrating Bender Jacks On

    Futurama1999

    Bender Jacks On

  3. Still from Wonka illustrating Paid in Chocolate

    Wonka2023

    Paid in Chocolate

  4. Still from Interstellar illustrating Mann Fakes the Telemetry

    Interstellar2014

    Mann Fakes the Telemetry

  5. Still from The Matrix illustrating Cypher Chooses the Steak

    The Matrix1999

    Cypher Chooses the Steak

  6. Still from Hot Fuzz illustrating Village of the Year

    Hot Fuzz2007

    Village of the Year

  7. Still from Naruto: Shippuden illustrating Infinite Tsukuyomi

    Naruto: Shippuden2015

    Infinite Tsukuyomi

  8. Still from Ocean’s Eleven illustrating The Fake Vault Feed

    Ocean’s Eleven2001

    The Fake Vault Feed

  9. Still from Spirited Away illustrating No-Face Buys Approval

    Spirited Away2001

    No-Face Buys Approval

  10. Still from Ferris Bueller's Day Off illustrating Attendance Record Hack

    Ferris Bueller's Day Off1986

    Attendance Record Hack

  11. Still from Memento illustrating Poisoning Future Memory

    Memento2001

    Poisoning Future Memory

  12. Still from Willy Wonka & the Chocolate Factory illustrating Wonka’s Computer Refuses the Bribe

    Willy Wonka & the Chocolate Factory1971

    Wonka’s Computer Refuses the Bribe

  13. Still from Family Guy illustrating Customer of the Week

    Family Guy2021

    Customer of the Week

  14. Still from True Detective illustrating Tuttle Captures Oversight

    True Detective2014

    Tuttle Captures Oversight

  15. Still from Family Guy illustrating Brian Rewinds the Gauge

    Family Guy2012

    Brian Rewinds the Gauge

  16. Still from True Detective illustrating Poisoning the Audit Trail

    True Detective2014

    Poisoning the Audit Trail

  17. Still from The Princess Bride illustrating Both Goblets Are Poisoned

    The Princess Bride1987

    Both Goblets Are Poisoned

  18. Still from Steins;Gate illustrating Operation Skuld Games the Observation

    Steins;Gate2011

    Operation Skuld Games the Observation

Relational index

Connected records

21