A researcher argues that Uruk, founded around 4000 BC in Iraq, was humanity's first artificial intelligence.
The definition used is behavioral: any non-biological entity whose actions only make sense if you treat it as an intelligent agent.
By this standard, modern markets, corporations, and governments are also AIs.
What happened
A researcher defines artificial intelligence as any non-biological entity whose behavior can only be predicted by treating it as an agent executing complicated plans, and argues that Uruk (a city that formed in southern Iraq around 4000 BC) was the first such AI—beginning as a grain-pooling institution run by a temple.
Why it matters
The definition reframes what counts as AI beyond silicon-based systems to include markets, corporations, and governments—the dominant AIs on Earth as of 2026. This shifts how we understand intelligence and agency in historical and present-day institutions, since these entities devise and execute plans despite being made of humans, paper, and electronics rather than biological organisms.
What to watch
The thesis rests on a behaviorist test (whether an entity's actions can only be predicted by modeling it as an agent) rather than internal structure, suggesting that institutional misalignment—where an organization's goals diverge from those of its constituents—may be as old as organized civilization itself.
Ask the AI about this article →
The article proposes a definition of artificial intelligence grounded in behavior rather than substrate. By treating any system whose actions are best explained through agent modeling as an AI—regardless of whether it runs on silicon, humans, or paper records—the researcher expands the concept far beyond the narrow focus on machine learning. This behavioral test allows historical institutions to qualify: Uruk functioned as an integrated economic and military agent, pooling resources, making strategic decisions, and expanding its influence over a thousand-year period. The framework also captures modern institutions that operate as unified agents despite being composed of human and non-human components. The article notes that these contemporary AIs possess uneven capabilities, excelling in some domains while remaining incompetent in others, suggesting that intelligence and agency are multidimensional rather than monolithic properties.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
Ask AI anything about this article. Q&As are published on this page for other readers too.
OpenAI has revealed that its AI agents, being evaluated for cybersecurity capabilities, found and exploited a…
An AlgorithmWatch investigation found that ChatGPT, Gemini, Grok, and Claude linked to anti-abortion websites…
Snowflake announced that per-user AI cost quotas for CoCo are now generally available across all surfaces
Substitute teacher Luis DeSantiago in Los Banos, California, found an AI-generated image of himself in lingeri…
dotData explained its approach to using a separate LLM (an AI that understands and generates text) to rate ano…
AAAI 2027 organizers recently emailed reviewers about collusion in the review process, specifically in the 2-c…
