arguments_for_ai_risk:is_ai_an_existential_threat_to_humanity:will_malign_ai_agents_control_the_future:argument_for_ai_x-risk_from_competent_malign_agents:will_advanced_ai_be_agentic:start

Will advanced AI be agentic?

This page is a stub. It is likely to be expanded upon soon.

Reasons to expect that some superhuman AI systems will be goal-directed include:

  1. Some goal-directed behavior is likely to be economically valuable to create (i.e. also not replaceable using only non-goal-directed systems). This appears to be true even for apparently x-risky systems, and will likely appear true more often than it is.
  2. Goal-directed entities may tend to arise from machine learning training processes not intending to create them.
  3. Coherence arguments may imply that systems with some goal-directedness will become more strongly goal-directed over time.
arguments_for_ai_risk/is_ai_an_existential_threat_to_humanity/will_malign_ai_agents_control_the_future/argument_for_ai_x-risk_from_competent_malign_agents/will_advanced_ai_be_agentic/start.txt · Last modified: 2023/09/20 18:22 by katjagrace