wireheading

English

Noun

wireheading (uncountable)

1. The use of direct brain interfaces.
   • 1992 January 24, David Wharton, “The New Edge : Magazines: The New Age meets the PC Age in Mondo 2000. The quarterly’s topics range from latest in high-tech products, jewelry and culture.”, in The Los Angeles Times:

     Mondo’s chaotic editorial blend ranges from stories on computer-chip jewelry to reports on computerized break-ins at automated-teller machines and “wireheading"--the implantation of electrodes in the brain.

2. The strategy of meeting goals by altering the perception of the current state rather than changing the state itself.
   • 2016 June 25, Tom Everitt, Marcus Hutter, “Avoiding Wireheading with Value Reinforcement Learning”, in Lecture Notes in Computer Science‎^[1], volume 9782, Springer, →DOI, pages 12–22:

     The constraint is defined in terms of the agent's belief distributions, and does not require an explicit specification of which actions constitute wireheading.

   • 2019, Stuart J. Russell, Human Compatible: Artificial Intelligence and the Problem of Control, Penguin, →ISBN, page 206:

     The tendency of animals to short-circuit normal behavior in favor of direct stimulation of their own reward system is called wireheading.