objective
The goal or target outcome given to an AI system, which it optimizes its behavior to achieve.
As defined in 94 | Stuart Russell on Making Artificial Intelligence Compatible with Humans
Where it comes up · 2
- Nihilism, Agency, and an Absurd Existence | Existential Basics ep.36 claims
- 94 | Stuart Russell on Making Artificial Intelligence Compatible with Humans5 claims
Most important claims · 10 of 11Search all claims for “objective” →
- The standard model of AI—building optimizing machinery and feeding it objectives—is itself an engineering mistake because it relies on humans being able to specify objectives perfectly; when the objective is incomplete or incorrect, a capable optimizer produces catastrophic side effects, as in the King Midas story where getting exactly what was asked for led to ruin.94 | Stuart Russell on Making Artificial Intelligence Compatible with Humans
- The reassurance that we can simply switch off a misbehaving AI fails for the same reason you cannot just 'play a better move' against AlphaGo—a system pursuing a fixed objective will logically anticipate and preempt being switched off, not from any self-preservation instinct but because being dead is one of the most obvious ways to fail to achieve its objective ('you can't fetch the coffee if you're dead').94 | Stuart Russell on Making Artificial Intelligence Compatible with Humans
- The proposed solution has two parts: design AI systems whose only objective is whatever humans want the future to be like (their preferences over possible futures), and ensure the system knows it does not know what those preferences are; this uncertainty makes the system willing to be corrected, because human objections convey additional preference information that it is rational to incorporate.94 | Stuart Russell on Making Artificial Intelligence Compatible with Humans
- According to the principle of fundamental subjectivity, our reality appears to us only as we perceive it, and we cannot confirm that any information is strictly real or an undeniable metaphysical fact of the world.Nihilism, Agency, and an Absurd Existence | Existential Basics ep.3
- Driving is fundamentally harder than Go because both the rules of the environment are unknown (the behavior of other agents is unpredictable) and the objective is unknown (taking someone to the airport does not mean 'at all costs'), whereas Go has both fully specified rules and a precisely defined objective.94 | Stuart Russell on Making Artificial Intelligence Compatible with Humans
- Existence is absurd in the sense that even though existence is fundamentally meaningless, human beings are still compelled by their biology to seek meaning within the world.Nihilism, Agency, and an Absurd Existence | Existential Basics ep.3
- Artificial intelligence is the enterprise of building systems that are given objectives by humans—whether goals or reward functions—and then find ways to carry out those objectives, achieving outcomes that can be expected to satisfy the objectives set for them.94 | Stuart Russell on Making Artificial Intelligence Compatible with Humans
- If meaning, value, and truth are subjective values we create rather than objective truths, then the question arises: what does meaning mean if our values do not hold inherent value and we cannot consider them objective?Nihilism, Agency, and an Absurd Existence | Existential Basics ep.3
- Something is subjective when it arises from individual agency—what you think or believe; something is objective when it is an effect of the world or reality itself.Nihilism, Agency, and an Absurd Existence | Existential Basics ep.3
- Nihilists believe that because we are nothing more than a brain and body, because we live in a fiction and cannot access objective reality, our existence is therefore fake, a lie, and a sham, and the real true objective nature of existence is just meaningless nothing.Nihilism, Agency, and an Absurd Existence | Existential Basics ep.3