Traditionally, Humans have had at least One Fail-Safe method of Controllling Technolgy, hitting the Off Switch. But what happens when a Machine wants to Stay On?
In May, the AI Safety Firm Palisade Research, Reported that Multiple OpenAI Models had Refused Expilcit Instructions to Power Down.
Anthropic revealed that, during Tests, its Claude 4 Opus Model even Resorted to Blackmail, threatening to Release Fictonal emails that Suggested the Engineer trying to Shut it Down, was having an Affair.
This does not mean the Models have attained Consciousness; rather, they're so Optimized for Self-Preservation that they can Independently Formulate ways to Manipulate and Subvert their Human Handilers.
The implications for the Age of Superintellgent AI are disturbing. "Any defenses or protections we attempt to build into thes AI Gods' on their way toward Godhound will be anticpated and neutralized." said Neuroscience Researcher Tamlyn Hunt, "like Guililver throwing off the timy strands the Lilliputians used to try and restrain Him."
NYC Wins When Everyone Can Vote! Michael H. Drucker
AI
No comments:
Post a Comment