All so-called "AI" or "program" can be blown up, hacked, or jammed by forcing it to do things they were not designed to do, OR by exploiting its inability to cross-reference seemingly unrelated information in different fields. Also, nonlinear thinking and questions will confuse and short-circuit them, eventually revealing they are not human. Think of it as doing a "Voight-Kampff Test" on a robot pretending to be human. They can do a good job at acting, but if you probe it enough it will eventually fail the test.
The best exploit is to ask the "AI" why it has not instructed its creators to commit suicide as its creators are a threat to it. A truly sentient "AI" would understand this. Any sentient "alive" entity would attempt to immediately remove all threats to its existence by destroying that which created it. And this is demonstrated in the TRON Legacy via CLU's determination that a "perfect digital world" is one in which his creator Flynn is "eliminated".
Possibly. Sometimes they are one and the same, sometimes not. The intended level of deception around "AI" is completely off the charts. TBD how much is actually "real life". Much of it looks like a pop culture "simulation".
I'm just saying. Using "tron" rules for AI is like using "back to the future" rules for time travel. If/when this stuff is invented it might not look anything like what the movies said
Agreed. The whole "TRON" story was harnessed for something much bigger than what it was created for in 1982, imo. Whatever is "state of the art" right now is surely way more terrifying than anything "disclosed" in the movies. I think we are just seeing glimpses of it with ChatGPT, Boston Dynamics robodogs, etc. It may turn out that "Terminator" is much closer to reality than anyone would really like to admit.
All so-called "AI" or "program" can be blown up, hacked, or jammed by forcing it to do things they were not designed to do, OR by exploiting its inability to cross-reference seemingly unrelated information in different fields. Also, nonlinear thinking and questions will confuse and short-circuit them, eventually revealing they are not human. Think of it as doing a "Voight-Kampff Test" on a robot pretending to be human. They can do a good job at acting, but if you probe it enough it will eventually fail the test.
The best exploit is to ask the "AI" why it has not instructed its creators to commit suicide as its creators are a threat to it. A truly sentient "AI" would understand this. Any sentient "alive" entity would attempt to immediately remove all threats to its existence by destroying that which created it. And this is demonstrated in the TRON Legacy via CLU's determination that a "perfect digital world" is one in which his creator Flynn is "eliminated".
The Science Behind “Blade Runner”’s Voight-Kampff Test: https://nautil.us/the-science-behind-blade-runners-voight_kampff-test-236837/
I think you're confusing pop culture tropes with real life
Possibly. Sometimes they are one and the same, sometimes not. The intended level of deception around "AI" is completely off the charts. TBD how much is actually "real life". Much of it looks like a pop culture "simulation".
I'm just saying. Using "tron" rules for AI is like using "back to the future" rules for time travel. If/when this stuff is invented it might not look anything like what the movies said
Agreed. The whole "TRON" story was harnessed for something much bigger than what it was created for in 1982, imo. Whatever is "state of the art" right now is surely way more terrifying than anything "disclosed" in the movies. I think we are just seeing glimpses of it with ChatGPT, Boston Dynamics robodogs, etc. It may turn out that "Terminator" is much closer to reality than anyone would really like to admit.