Term of the Moment
accelerator card
Redirected from: AI threats
Definition: AI anxiety
(1) (Artificial Intelligence anxiety) The fear that computers will replace people and jobs in the future. As AI becomes more sophisticated, it is able to replicate more human decision making. This form of automation anxiety dates back to the late 1700s when mechanical looms replaced jobs in the textile industry (see Jacquard loom).
(2) (Artificial Intelligence anxiety) The fear that computers will make future decisions that will be harmful to humanity. Many people, including famous personalities such as Bill Gates, Elon Musk and Stephen Hawking, have expressed concern that society might be irreparably damaged as computers increasingly replace human decision making. Of course, one would hope that the decision to launch a nuclear weapon would never be made by an algorithm. However, it is possible to envision that under the rationale of "eliminating human emotion," this might eventually come to pass. See GPT, technology singularity, robot and AI.
A Remarkable Observation
Sooner Than You Think
Anxiety in the 1940s
A Warning
(2) (Artificial Intelligence anxiety) The fear that computers will make future decisions that will be harmful to humanity. Many people, including famous personalities such as Bill Gates, Elon Musk and Stephen Hawking, have expressed concern that society might be irreparably damaged as computers increasingly replace human decision making. Of course, one would hope that the decision to launch a nuclear weapon would never be made by an algorithm. However, it is possible to envision that under the rationale of "eliminating human emotion," this might eventually come to pass. See GPT, technology singularity, robot and AI.
Best-selling author and history professor at the Hebrew University in Jerusalem, Yuval Noah Harari has a lot to say about AI. One of his observations in the October 2024 Rich Roll Podcast "Our AI future is way worse than you think" is that as time goes on, instead of feeding human-generated data into an AI model for training, the output of previous AI systems will be fed back into new AI models for training. In such cases, the AI systems will be learning from themselves and not human experiences.
In 2023, best-selling author and ex-chief business officer for Google's X research division, Mohammad "Mo" Gawdat said in Steven Bartlett's "Diary of a CEO" podcast that AI will become much more intelligent than humans in only a few months. Gawdat claimed that in a couple years, Bartlett will be interviewing a robot for his podcast, and in 10 years, we will be hiding from the machines. Needless to say, Gawdat's comments caused enormous controversy.
The above quote appeared in 1947 in an essay on social criticism entitled "Dialectic of Enlightenment" by German philosophers Max Horkheimer and Theodor W. Adorno. The quote was referenced in "The Loop" by Jacob Ward (see below).
In his 2022 book, science and tech journalist Jacob Ward says our innermost loop is the behavior we inherited, but the outermost loop is the way technology, capitalism, marketing and politics use AI to sample our behavior and reflect those patterns back at us. He claims the real threat is what happens over time as AI alters future behavior.