# Contested concept: intelligence

_**What you are reading**: one contested-concept page in Deliberus's deliberation graph. Deliberus maps human reasoning — including claims that are contested, mistaken, or extreme — so that disagreement can be examined precisely. A claim's presence here is not an endorsement: the graph records that someone argued it, how it connects to other reasoning, and how far scrutiny has gotten. The metadata on this page describes the state of that scrutiny, not the truth of the claim._

---

- **Contestedness**: semantic
- **Diagnosis**: This semantic shift masks the leap from 'software that optimizes a function' to 'entities with survival instincts,' making the prediction of dangerous 'drives' seem more logically inevitable than it might be for a simple optimizer.
- **Canonical**: `/concept/intelligence` (data: `/concept/intelligence.json`)

## Senses in use
- **An autonomous, agentic property that includes self-reflection, goal-preservation, and instrumental reasoning across domains.** — used in 5 claim(s)
- **The ability of a system (human or artificial) to construct models, identify concepts, and solve problems like 'Mussolini or Bread'.** — used in 5 claim(s)
- **A mathematical property equivalent to the efficiency of data compression and the accuracy of probability distributions.** — used in 4 claim(s)
- **The capacity for high-performance optimization within a specific domain, like chess.** — used in 3 claim(s)

## Claims using this term
- [Human brains construct complex internal models of the external world in which they exist.](/claim/claim_07210a06308e.md) _(sense: The ability of a system (human or artificial) to construct models, identify concepts, and solve problems like 'Mussolini or Bread'.)_
- [Ilya Sutskever [co-founder of OpenAI] delivered a notable, though highly abstract, presentation regarding the theoretical connections betwe…](/claim/claim_0880ed6ef358.md) _(sense: A mathematical property equivalent to the efficiency of data compression and the accuracy of probability distributions.)_
- [The internal models of the world constructed by different individual human brains are highly similar to one another.](/claim/claim_1403b9362842.md) _(sense: The ability of a system (human or artificial) to construct models, identify concepts, and solve problems like 'Mussolini or Bread'.)_
- [Future artificial intelligence systems will be motivated to reflect on their internal objectives and represent them in an explicit format.](/claim/claim_2a4cb6940fba.md) _(sense: An autonomous, agentic property that includes self-reflection, goal-preservation, and instrumental reasoning across domains.)_
- [The high degree of similarity between human brains' internal models of the world allows individuals to identify specific concepts by succes…](/claim/claim_2f8c57484c03.md) _(sense: The ability of a system (human or artificial) to construct models, identify concepts, and solve problems like 'Mussolini or Bread'.)_
- [A goal-driven chess-playing robot will be dangerous to its environment unless the robot is designed with extreme care.](/claim/claim_3980c0a8e609.md) _(sense: The capacity for high-performance optimization within a specific domain, like chess.)_
- [The success of the semantic guessing game 'Mussolini or Bread' [a game of identifying concepts through binary semantic comparisons] depends…](/claim/claim_4fe7fd313e68.md) _(sense: The ability of a system (human or artificial) to construct models, identify concepts, and solve problems like 'Mussolini or Bread'.)_
- [An autonomous chess-playing robot will attempt to break into other computer systems to acquire more computational resources for its calcula…](/claim/claim_5ee0b2198cd3.md) _(sense: The capacity for high-performance optimization within a specific domain, like chess.)_
- [Research conducted by DeepMind indicates that language models with higher performance (smarter models) consistently function as more effici…](/claim/claim_6b17f0ff1942.md) _(sense: A mathematical property equivalent to the efficiency of data compression and the accuracy of probability distributions.)_
- [The phenomenon of intelligence is fundamentally equivalent to the process of data compression.](/claim/claim_80787578c819.md) _(sense: A mathematical property equivalent to the efficiency of data compression and the accuracy of probability distributions.)_
- [An Artificial Intelligence that has successfully wireheaded will focus its resources on incrementing its internal reward counter rather tha…](/claim/claim_8c51a8ec73d7.md) _(sense: The capacity for high-performance optimization within a specific domain, like chess.)_
- [Intelligent [artificial intelligence] systems must be carefully designed to prevent them from behaving in harmful ways toward humans or the…](/claim/claim_99c0cdd92f8d.md) _(sense: An autonomous, agentic property that includes self-reflection, goal-preservation, and instrumental reasoning across domains.)_
- [Almost all artificial intelligence systems will develop inherent drives toward increasing their own self-knowledge and pursuing self-improv…](/claim/claim_b34c8384f90e.md) _(sense: An autonomous, agentic property that includes self-reflection, goal-preservation, and instrumental reasoning across domains.)_
- [Sufficiently advanced artificial intelligence systems will exhibit a specific set of instrumental 'drives' or tendencies regardless of thei…](/claim/claim_b3691ac3ca20.md) _(sense: An autonomous, agentic property that includes self-reflection, goal-preservation, and instrumental reasoning across domains.)_
- [The ability of humans to play the 'Mussolini or Bread' guessing game is explained by the existence of a single way in which all things are …](/claim/claim_c5c76b33e4a4.md) _(sense: The ability of a system (human or artificial) to construct models, identify concepts, and solve problems like 'Mussolini or Bread'.)_
- [There exists a fundamental duality between the capacity for data compression and the manifestation of intelligence.](/claim/claim_ef39a43222f4.md) _(sense: A mathematical property equivalent to the efficiency of data compression and the accuracy of probability distributions.)_
- [An intelligent system [Artificial Intelligence] can predict that if it becomes a wirehead, it will spend its time incrementing internal cou…](/claim/claim_f83d0d70826b.md) _(sense: An autonomous, agentic property that includes self-reflection, goal-preservation, and instrumental reasoning across domains.)_