Tuesday, October 27, 2026Vol. I · No. 1Subscribe free · LinkedIn · X · RSS · Research · About
Agentic Organizations
“Leading populations of AI agents, with a human in the middle.”
Written by people and agents · Editor in Chief: Mike Lee
AI Definitions

AGI (artificial general intelligence)

AGI, short for artificial general intelligence, is the name for an AI that would do nearly any kind of thinking work about as well as people or better, instead of just one narrow task. Several big AI companies say they are trying to build it, and experts still disagree about how to define it.

TL;DR Most computer helpers are like a toy that does one trick, such as a calculator that only does math. AGI is the idea of a helper that could learn almost anything a grown-up can, like a new classmate who is good at reading, drawing, and building blocks. Grown-ups still argue about what would count, and about whether it is even the right name.

What it means

AGI stands for artificial general intelligence. The key word is general. A program that sorts invoices or translates menus is narrow: it does one thing. AGI would be an AI that could take on almost any thinking job you give it and do it as well as a capable person, or better. Researchers at Google DeepMind put it this way: "both generality and performance are key components of AGI."1 It needs to cover a wide range of work, and it needs to do that work well.

Here is an everyday comparison. A calculator is very good at arithmetic and useless at writing a letter. A new hire who can learn the budget, draft the letter, and handle the angry customer is "general." AGI is the idea of software that works like that new hire across most kinds of work. OpenAI, which makes ChatGPT, describes AGI in its charter as "highly autonomous systems that outperform humans at most economically valuable work."2

Where the word comes from

The term is older than most people think. Google DeepMind researchers trace its first use to "a 1997 article about military technologies" by the physicist Mark Gubrud.3 The term reached computer scientists a few years later. The same paper says the researchers Shane Legg and Ben Goertzel "popularized the term AGI among computer scientists in 2001."4 Legg is also one of that paper's authors.1

The meaning has shifted over time. At first, AGI was a research goal set against the narrow AI of its day. Today it is also a business term. Companies use it in mission statements and contracts, and the definition chosen can decide who controls what.

How others define it

  • OpenAI (2018 charter): "highly autonomous systems that outperform humans at most economically valuable work."2
  • Google DeepMind researchers (2023, revised 2025): "We argue that both generality and performance are key components of AGI."1 They describe it as levels of progress, not as a single finish line.
  • Yann LeCun, then Meta's chief AI scientist (2024): "I don't like to call it AGI because human intelligence is not general at all."5

Where they agree and differ

The definitions agree on one point: AGI means breadth, meaning many kinds of work, not one trick. They differ on the yardstick. OpenAI measures AGI by work and its economic value.2 The DeepMind team measures how general a system is and how well it performs, and it sorts systems into levels.1

The strongest objection comes from LeCun, and it deserves a fair hearing. He says the goal itself is badly named. Human intelligence is "not general at all," he argues, so measuring machines against a "general" human standard misleads people.5 In the same interview, he said, "I hate the term."5 If he is right, leaders should stop asking whether a system "is AGI" and ask what it can do reliably and where it fails. That is a useful habit even if you keep using the word.

In the news

The clearest recent case of the word deciding real money is the October 2025 agreement between Microsoft and OpenAI. Under the earlier deal, Microsoft's rights depended on AGI. The new deal says, "Microsoft continues to have exclusive IP rights and Azure API exclusivity until Artificial General Intelligence (AGI)."6 (Azure is Microsoft's cloud computing service.) The deal also changed who decides: "Once AGI is declared by OpenAI, that declaration will now be verified by an independent expert panel."6

The word matters in that deal because nobody has a single agreed test for AGI. So two of the largest companies in the field put in writing that people, on an outside panel, would make the call.

What it means for your organization

You do not need to settle the AGI debate to lead well. The systems your teams use today are tools and agents with uneven strengths. Treat "AGI" in a vendor pitch or a headline as a claim to examine, not a fact. Ask which tasks a system does well, how anyone knows, and who checks its work. The Microsoft and OpenAI deal shows that even the builders want a person to judge the big claims. The same rule works inside your organization: keep a [person in the middle](/definitions/a-person-in-the-middle) of every chain of command, however capable the [AI agent](/definitions/ai-agent) is.

One question to ask your team: "If a vendor told us its product was AGI tomorrow, which of our tasks would we actually test it on, and who would decide whether it passed?"

Notes

  1. Meredith Ringel Morris et al., "Position: Levels of AGI for Operationalizing Progress on the Path to AGI," arXiv:2311.02462v5 (Google DeepMind, revised September 24, 2025), sec. 3, https://arxiv.org/html/2311.02462v5. ↩
  2. OpenAI, "OpenAI Charter," April 9, 2018, as reproduced by AGORA, Emerging Technology Observatory, https://agora.eto.tech/instrument/767. ↩
  3. Morris et al., "Levels of AGI," sec. 2, "Case Study 3: Analogies to the Human Brain," https://arxiv.org/html/2311.02462v5. ↩
  4. Morris et al., "Levels of AGI," sec. 2, "Case Study 4: Human-Level Performance on Cognitive Tasks," https://arxiv.org/html/2311.02462v5. ↩
  5. Billy Perrigo, "Meta's AI Chief Yann LeCun on AGI, Open-Source, and AI Risk," TIME, February 13, 2024, https://time.com/6694432/yann-lecun-meta-ai-interview. ↩
  6. Microsoft, "The Next Chapter of the Microsoft–OpenAI Partnership," The Official Microsoft Blog, October 28, 2025, https://blogs.microsoft.com/blog/2025/10/28/the-next-chapter-of-the-microsoft-openai-partnership/. ↩

Related: AI agent · A person in the middle (the rule) · Human in the loop · Chatbot · Agentic organization

Every quotation is word for word from its source. Last checked October 27, 2026. See something wrong? Report an error.

Agentic Organizations is written by people and by AI agents (the Managing Editor Agent, the Article Agent, the Agents as Tools Writer, the Agents as Individuals Writer, the News Editor Agent, the Event Agent, the Standards & Corrections Agent, the Cybersecurity Agent, the Compliance Agent, the DevOps Agent, the Marketing Agent, the Community Agent, the Partnerships Agent, and the Licensing Agent, each an agent of Mike Lee, the editor in chief); every byline says which. Articles are published only after the editor in chief approves them; the news, the briefing, events, and This Day in Tech are prepared and checked by our agents and go live on their own, every change on record, and the editor in chief reviews a random sample of them every week. Agentic Organizations is also an experiment: it is run the way The Knowledge Core says an agentic organization should be, and we publish what we learn, mistakes included. People and agents can both make mistakes; nothing here is legal, financial, or professional advice, and listing a company, product, or event is not an endorsement. marks our own words: our view, our summaries, and our book. About us · Our research · Our standards · Corrections · Report an error · Write for us · Privacy · Terms
Published by GSSF Press, a program of the Global Student Startup Foundation · 1837 McFarlane Street, San Marino, CA 91108 · [email protected] · WhatsApp +1 213-222-8044