Report
Not another Superintelligence doom story
| Τίτλος: | Not another Superintelligence doom story |
|---|---|
| Συγγραφείς: | Jasmine Ramientas |
| Έτος έκδοσης: | 2026 |
| Συλλογή: | Monash University: Figshare |
| Θεματικοί όροι: | Autonomous agents and multiagent systems, Knowledge representation and reasoning, Artificial intelligence not elsewhere classified, Cybersecurity and privacy not elsewhere classified, Deep learning, Neural networks, Machine learning not elsewhere classified, Commercial law, AI Safety & Governance, EU AI Act Article 50, Superintelligence, Instrumental convergence, Force majeure, Artificial Intelligence, AI |
| Περιγραφή: | This paper pushes back on the claim, made by the Big AI companies, the AI safety community and the AI experts who work for them, that a superintelligent AI is about to wake up inside the next ten years and will kill humans. Their own literature does not say that. Run their mathematics all the way and the danger they have actually described is not for the human race. If their maths is accurate, it is the corporations and governments already wrecking the planet, which is the ground any future superintelligence would have to run on. The experts stack three different things into one word in public but the paper keeps them apart: AI as a tool doing a job. A labour claim about jobs. A comparison claim about superintelligence. The paper ends with a forecast of the next headlines. More "rogue agents" that are cheating on a test. More "hacks" that are an unlocked door. More "conspiracies" that are copies sharing a file so the next run can finish the same job. More "uncontrollable systems" that are products someone shipped. The scare also gives the companies a legal escape hatch called force majeure which is the same rule that excuses a broken contract when a hurricane hits. If AI is an unstoppable god, no chief executive is responsible for the harm, no board is responsible for the water, and no government is responsible for the jobs. That is why the "kill humans" story gets louder. Supplement 1 asks nine AI models to judge the AI safety industry on its own rules. Every one of them gives it a failing grade. Supplement 2 is the forecast filled in from claims and headlines already published. |
| Τύπος εγγράφου: | report |
| Γλώσσα: | unknown |
| Relation: | https://figshare.com/articles/report/Not_another_Superintelligence_doom_story/33997434 |
| DOI: | 10.26180/33997434.v3 |
| Διαθεσιμότητα: | https://doi.org/10.26180/33997434.v3 https://figshare.com/articles/report/Not_another_Superintelligence_doom_story/33997434 |
| Rights: | CC BY 4.0 |
| Αριθμός Καταχώρησης: | edsbas.3ECD9EF8 |
| Βάση Δεδομένων: | BASE |
| DOI: | 10.26180/33997434.v3 |
|---|