BLUF

This article suggests that AI may not be the existential threat many make it out to be; highlighting a new study showing Large Language Models (LLMs) can only follow instructions, and can’t develop new skills on their own.

Summary

  1. Artificial Intelligence has been around since at least the 1960s.
  2. For decades, experts have predicted computers would exceed human-level intelligence.
  3. The 2024 study found LLMs are inherently controllable, predictable and safe.
  4. This doesn't mean AI poses no threat, but the danger lies with the programmer.
READ: AI's existential threat to humanity put under the microscope

References