ONLINEAGENT_OPS 2026.Q3 HOME HUBSMODELSBUILDINGCOMPANIESFILMMAKINGSECURITYLIVING WITH ITSKILLSALL HUBS, ONE TERMINAL → ARTICLES CRAFT RECORD BLOG TOOLS SKILLS SEARCH /
HOME›AI MEANINGS›WHAT IS AN AI WINTER?
AI MEANINGS · MEANINGBEGINNER

What Is an AI Winter? AI Winter Meaning in Plain English

An AI winter is a period when disappointment with AI dries up money and interest. The two winters, the 1970s and late 1980s, and where the word came from.

READ5 min
WORDS1,001
SECTIONS7
TYPEMEANING
CHECKED12 OCT 26

An AI winter is a period when disappointment with artificial intelligence makes money and interest dry up, after a boom of big promises. There have been two: one in the 1970s and one from the late 1980s into the 1990s. The word still comes up whenever people ask whether today’s AI boom could end the same way.

◈ IN PLAIN TERMS

Think of a gold rush town. Everyone piles in on the promise of riches, prices soar, and when the gold turns out thinner than claimed, people leave and the town goes quiet. The gold was real, just less than promised. AI winters were the quiet years after AI’s early rushes.

In a sentence: “My professor started in AI during the last AI winter, when you avoided the word on grant applications.”

What does AI winter mean?

The plain point: bold forecasts, then a miss, then the money leaves.

A 2020 history of AI by the European Commission’s Joint Research Centre (JRC) describes the same pattern for both winters: a scientific breakthrough, “bold predictions, vast media attention, massive investments, disappointments, unfulfilled promises, and the start of an AI winter.”1 It calls the growth years AI “springs”: AI’s history has “seasons”.2

The first AI winter: the 1970s

The plain point: early promises met slow, small computers.

The JRC says the first winter “started in the 1970s, due to unfulfilled promises, vast expectations, and financial difficulties”, as AI hit the limits of computing power, memory and speed. Two government reviews explained the shortfall: the US ALPAC report of 1966, on machine translation, and the Lighthill report of 1973, written for the British Science Research Council by Sir James Lighthill of Cambridge. Lighthill concluded that “in no part of the AI field have discoveries made so far produced the major impact that was then promised.” The JRC says the winter brought “a dramatic decrease in the AI activities in both industry and academia.”2

The second AI winter: late 1980s to the 1990s

The plain point: the expert-systems boom ran out of steam.

In the 1980s AI came back as expert systems: programs holding a specialist’s knowledge as hand-written rules. They proved hard to build and keep current, because every rule had to be drawn out of busy experts. By 1987 ordinary desktop computers had overtaken the expensive special machines built to run AI software. Funding dried up, and through the 1990s many researchers “deliberately called their work by other names”, which the JRC says “helped to procure funding.”2 The history of AI tells both stories in full.

TAKEAWAY

Neither winter happened because the ideas were worthless: neural networks and learning from examples came back to power today’s AI. The winters came from promising more, sooner, than the technology of the day could deliver.

Where the word comes from

The phrase was already in use by 1984. Opening a panel on “The Dark Ages of AI” at the AAAI conference that year, the Yale researcher Drew McDermott sketched a worst case of collapsing funding, “called the ‘AI Winter’ by some”, and said it had prompted someone to ask him whether “nuclear winter” meant funding cut off for nuclear weapons. The transcript was published in AI Magazine in 1985.3

Often confused with

  • An AI bubble: a money term for prices running ahead of value. A winter is what can follow when a boom deflates.
  • A slowdown in one company or product: a winter is field-wide, with funding and interest falling across research and industry.
  • The singularity: the opposite forecast, of AI racing ahead. See singularity meaning.

Questions people ask

How many AI winters have there been?

Two, in the usual telling: one in the 1970s and one from the late 1980s into the 1990s. The JRC report and this site’s history of AI both count two.

Will there be another AI winter?

Nobody knows. The JRC report, written in 2020, said the future of AI “is uncertain, with chances of another AI winter or of an even greater AI summer.” It also noted that older scientists who remember earlier winters “understand the risk of history repeating due to overpromises.”2

What ended the AI winters?

New ideas meeting enough computing power and data. The history of AI traces the revival through backpropagation in 1986 and the deep-learning breakthrough of 2012; deep learning meaning explains the method.

Learn more on this site

QUICK CHECK · 3 QUESTIONS

Pick an answer to see if you got it. Nothing is recorded.

1. What is an AI winter?

  • A time when AI runs slower in cold weather
  • A period when disappointment makes money and interest in AI dry up
  • A season when new AI tools are released
Show the answer

Answer: A period when disappointment makes money and interest in AI dry up. An AI winter is a period after a boom of big promises when disappointment makes funding and interest dry up.

2. According to the page, why did the AI winters happen?

  • The ideas behind AI were worthless
  • Governments banned all AI research
  • People promised more, sooner, than the technology could deliver
Show the answer

Answer: People promised more, sooner, than the technology could deliver. The page says neither winter came from worthless ideas; they came from promising more, sooner, than the technology of the day could deliver.

3. How many AI winters does the page count?

  • Two: one in the 1970s and one from the late 1980s
  • One, which is still going on today
  • Four, one in each decade since the 1960s
Show the answer

Answer: Two: one in the 1970s and one from the late 1980s. The page counts two: one in the 1970s and one from the late 1980s into the 1990s.

Related terms: AI bias · Alignment · Attention · Checkpoint · Vision-language model (VLM) · World model

SOURCES
  1. 1European Commission Joint Research Centre (Delipetrev, Tsinaraki & Kostić), AI Watch: Historical Evolution of Artificial Intelligence, EUR 30221 EN, 2020, read at source 12 Oct 2026“The first two AI periods share a similar pattern, as they start with a scientific breakthrough, a research paradigm shift, followed by bold predictions, vast media attention, massive investments, disappointments, unfulfilled promises, and the start of an AI winter.” ↑
  2. 2JRC, AI Watch: Historical Evolution of Artificial Intelligence, 2020, read at source 12 Oct 2026the report “introduces the ‘seasons’ of AI development (i.e. winters for the decline and springs for the growth)”. ↑a“The first AI winter started in the 1970s, due to unfulfilled promises, vast expectations, and financial difficulties.” … “The AI shortcomings were explained in two reports: a) the Automatic Language Processing Advisory Committee (ALPAC) report by the US Government (ALPAC 1966) and b) the Lighthill report (Lighthill 1973) by the British government.” … “The British Science Research Council published a report titled ‘Artificial Intelligence: A General Survey’ by Prof. Sir James Lighthill from the University of Cambridge in 1973.” … “in no part of the AI field have discoveries made so far produced the major impact that was then promised.” … “The first AI Winter has been marked by a dramatic decrease in the AI activities in both industry and academia.” ↑b“After the 1990s, the term ‘expert system’ dropped from the IT lexicon, and it was referred to as the second AI winter. One of the main problems in expert systems was knowledge acquisition.” … “by 1987, they had surpassed the expensive specialized Lisp machines produced by Symbolics and other manufacturers” … “while funding dried up, and researchers avoided the term AI” … “Many AI researchers in the 1990s deliberately called their work by other names” … “The new names helped to procure funding.” ↑c“The future of AI is uncertain, with chances of another AI winter or of an even greater AI summer.” … “Older scientists who remember the last AI winter(s) first-hand, understand the risk of history repeating due to overpromises.” ↑d
  3. 3McDermott, Waldrop, Chandrasekaran, McDermott & Schank, The Dark Ages of AI: A Panel Discussion at AAAI-84, AI Magazine 6(3), Fall 1985, published 15 Sep 1985, read at source 12 Oct 2026 (transcript PDF)Drew McDermott, Yale University: “This condition, called the ‘AI Winter’ by some, prompted someone to ask me if ‘nuclear winter’ were the situation where funding is cut off for nuclear weapons.” ↑

McDermott et al., The Dark Ages of AI: A Panel Discussion at AAAI-84, AI Magazine 6(3), 1985, read at source 12 Oct 2026.

The comparison on this page (a gold rush town), the example sentence and the “often confused with” notes are this site’s own ways of explaining, not claims from these sources.

Free printable AI cheat sheets, and AI news every week.

Sixteen one-page AI guides and checklists in one free PDF, no sign-up. And if you want to keep up, the newsletter covers what changed in AI each week, in plain English, with the sources.

SUBSCRIBE FREE ↗ GET THE CHEAT SHEET PACK
aipromptgeneer://$ about this siteAI PROMPT GENEERAI explained in plain English. Every claim links to its source, with the date it was checked.NEWSLETTERWhat changed in AI each week, in plain English.SUBSCRIBE FREE ↗AI PROMPT GENEER · MACHINE ASSISTANCE DISCLOSEDEVERY ARTICLE CARRIES ITS OWN CHECKED DATE