Classical computer science established that internal world models are sufficient for optimal performance, but proving they are strictly necessary under realistic conditions remained an unresolved challenge. A new study presented at UAI 2026 resolves this gap by framing prediction as a betting decision to show how minimizing regret forces agents to retain past observations. Without these preserved predictive distinctions, an autonomous agent fails to tell apart currently identical situations and inevitably suffers poor performance. Consequently, competent agents must at minimum maintain decision-relevant knowledge regarding the observable consequences of their actions.
Key points
Agents operating in partially observed environments must preserve past predictive distinc…: Agents operating in partially observed environments must preserve past predictive distinctions to avoid sacrificing operational performance and accumulating regret.
Competent autonomous agents must strictly retain decision-relevant knowledge regarding th…: Competent autonomous agents must strictly retain decision-relevant knowledge regarding the observable outcomes that follow their actions.
Researchers can now deduce the internal predictive structures an agent must possess solel…: Researchers can now deduce the internal predictive structures an agent must possess solely by evaluating its task competence.
Sources checked by our analysis pipeline: 1
Part I: Capable Agents
As AI agents swiftly become increasingly capable and ubiquitous throughout society, it becomes a more pressing matter to understand (1) what we might expect to emerge inside of them, and (2) what, if any, guarantees we may have in controlling them.
My long-term research vision is to build a natural science of autonomous intelligence: executable theories of how agents generate useful behavior in open worlds, and how those mechanisms can remain under human control. In Part I of this series, I will start with what capable agents must necessarily have inside of them. Which properties of their internals must necessarily follow from observing their behavior of doing a task well, even if we did not explicitly put them into the architecture?
What competence makes necessary
Classical results in control and reinforcement learning show that optimal behavior can be implemented using belief states and a model of the environment: see Smallwood and Sondik (1973) and Kaelbling, Littman, and Cassandra (1998). These are sufficiency results — they do not establish necessity. Could a completely different implementation do just…
The privacy authority AP conducted an investigation into six different hardware models ranging from under €100 to over €500, discovering critical safety and cybersecurity defects across the products. Testers documented dangerous misinformation during health scenarios, alongside privacy vulnerabilities that permit unauthorized camera access and undetected recording. These failures pose acute hazards for vulnerable groups, including visually impaired individuals who may depend on the technology to navigate their sur…
Strong investor demand is propelling the artificial intelligence company toward a $1.9 trillion valuation, defying an ongoing selloff across the wider technology sector. Backed by substantial compute alliances with prominent cloud providers, the firm secured high-profile coverage from major brokerages anticipating explosive long-term revenue growth.
Defensive cybersecurity efforts are facing critical friction when commercial artificial intelligence guardrails prevent legitimate developers from testing and patching their own systems. Anthropic's restrictions forced an engineer to navigate burdensome exemption processes that mandate identity disclosure and surrender intellectual property safeguards. Consequently, defenders are actively abandoning restricted proprietary platforms for open-weights alternatives to complete routine vulnerability remediation.