CD Consulting R&D

CD Consulting R&D · KB-RISK · entry 1 · compiled 2026-09-07 from the seven public WEF Global Risks Reports · published 7 September 2026

Technology, Information and AI Risks in the WEF Global Risks Reports, 2020–2026

A seven-edition comparative analysis of the World Economic Forum annual summary sections

If you brief a board on risk, this is where AI entered the list, and how fast.

Compiled 2026-09-07 · Sources: the seven WEF Global Risks Reports (editions 2020–2026)

Contents

1. Scope and method
2. The measurement problem: three incompatible regimes
3. What counts as a technology or information risk
4. Year by year, 2020–2026
5. The comparable series, 2023–2026
6. Four cross-cutting movements
7. Artificial intelligence: the arrival of a category
8. The short-term / long-term gap, and what it means
9. What the WEF does not say
10. Findings in brief
11. Source table

1. Scope and method

This analysis reads only the annual summary sections of each report — the Preface, Key Findings / Executive Summary, and the Global Risks Landscape figures that carry the ranked outputs of the Global Risks Perception Survey (GRPS). Where the summary points to an in-depth chapter on a technology, information or AI theme, that chapter is used to explain the ranking, not to substitute for it.

The extraction is restricted to:

• every risk the WEF itself files under the Technological category, in whatever wording that year's list used; and
• information-integrity risks wherever the WEF placed them — this matters, because misinformation has moved between categories over the period and is filed as Technological from 2023 onwards.

All ranks quoted are the WEF's own published ranks. Where a figure gives severity scores on the 1–7 Likert scale, those are reproduced; where the report does not publish a score, none is invented. The underlying rank data is in the companion workbook WEF-GRR-2020-2026-tech-info-AI-risk-rankings.xlsx, same folder.

2. The measurement problem: three incompatible regimes

Any year-on-year comparison of these reports has to survive the fact that the WEF changed its central question twice in seven years. Reading a 2021 rank against a 2025 rank as if they measured the same thing produces nonsense. The three regimes are:

Two further discontinuities compound this:

• The risk list itself changes. 30 risks (2020) → 35 (2021) → 37 (2022) → 32 (2023) → 34 (2024) → 33 (2025) → 33 (2026). A risk falling in rank may simply be a longer or shorter list.
• Risk names change, and with them their boundaries. Cyber insecurity (2024) becomes Cyber espionage and warfare (2025) — a narrower, more state-centric definition — then reverts to Cyber insecurity (2026). Technological power concentration (2024) disappears in 2025, partly absorbed by an economic risk, Concentration of strategic resources and technologies. Online harms is introduced in 2025. Most consequentially, there is no stand-alone AI risk before the 2024 edition.

Consequence for this analysis. The 2023–2026 editions form a genuinely comparable four-point series and carry the weight of the quantitative argument. 2020–2022 are read qualitatively, as the prehistory that makes the 2024 break legible.

3. What counts as a technology or information risk

The technological category, year by year, as the WEF defined it:

Three structural observations follow.

First, the category is remade in 2023–2024, not extended. Between the 2022 and 2024 lists, four of six technological risks are replaced. The vocabulary shifts from infrastructure and access (IT infrastructure breakdown, digital inequality, digital power concentration) to content, control and cognition (misinformation, censorship and surveillance, online harms, AI). This is a change in what the WEF believes technology threatens: no longer chiefly the machinery, but the information environment and the people in it.

Second, digital inequality is written out. Digital inequality and lack of access to digital services — a top-10 likelihood risk in 2021 — has no successor after 2023. The distributional question does not disappear from the reports; it migrates into the AI narrative (the 2026 report notes AI adoption at 27% of the working-age population in North America against 9% in Sub-Saharan Africa) but it no longer has a risk of its own to be ranked. Anything not on the list cannot be measured.

Third, misinformation becomes technological. In the 2021–2022 regime the nearest equivalents — Backlash against science, Erosion of social cohesion — sat under Societal. From 2023 onward, Misinformation and disinformation is a technological risk. That reclassification is itself an analytical claim: that the driver is the tooling, not the polity.

4. Year by year, 2020–2026

2020 — technology as an infrastructure and governance problem

The GRPS 2019–2020 puts two technological risks in the top 10 by likelihood — Data fraud or theft (#6) and Cyberattacks (#7) — and two by impact: Information infrastructure breakdown (#6) and Cyberattacks (#8). This is the high-water mark for cyber risk in the whole period as a share of the top 10.

The dedicated chapter, Wild Wide Web: Consequences of Digital Fragmentation, frames the issue as a governance vacuum around Fourth Industrial Revolution technologies. AI appears in it, but as an upside — "artificial intelligence (AI) alone is expected to boost global growth by 14% by 2030" — set beside 5G, quantum computing and IoT. The risk named is not what these technologies do, but that "the current lack of global technology governance and the presence of cybersecurity blind spots increase the risk of a fragmented cyberspace and competing technology regulations."

The stand-alone risk Adverse consequences of technological advances is defined as "intended or unintended adverse consequences of technological advances such as artificial intelligence, geo-engineering and synthetic biology" — AI is named in the definition, four years before it gets a risk of its own. It does not reach either top 10.

2021 — the pandemic reframes technology as dependency and divide

The GRPS 2020 is fielded during COVID-19, and technological risk is read through it. Two technological risks enter the 10-year likelihood top 10: Digital power concentration (#6) and Digital inequality (#7). Neither reaches the impact top 10.

More revealing is the new Global Risks Horizon question, which sorts risks by when they become a critical threat:

• 0–2 years ("clear and present dangers"): Cybersecurity failure #4 (39.0%), Digital inequality #5 (38.3%)
• 3–5 years ("knock-on effects"): IT infrastructure breakdown #2 (53.3%), Cybersecurity failure #8 (49.0%), Tech governance failure #9 (48.1%)
• 5–10 years ("existential threats"): Adverse tech advances #4 (50.2%)

This is the first time the reports place an undifferentiated frontier-technology risk in an explicitly existential long-term band, ranked immediately behind weapons of mass destruction, state collapse and biodiversity loss. The long-horizon anxiety that AI will later carry is already present — it simply has no name yet.

2022 — the retreat, and the WEF's own warning about it

The 2022 edition marks the low point for technological risk. In the horizon question, Cybersecurity failure falls to #7 at 0–2 years (19.5%) and #8 at 2–5 years (14.6%); Digital inequality to #9 (18.2%). Adverse tech advances falls from #4 to #8 in the 5–10 year band (14.9%).

Decisively, in the new "most severe risks on a global scale over the next 10 years" ranking, no technological risk appears in the top 10 at all — it is climate, extreme weather, biodiversity, social cohesion, livelihoods, disease, debt and geoeconomic confrontation.

The report says so itself, and names the consequence:

"Technological risks — such as 'digital inequality' and 'cybersecurity failure' — are other critical short- and medium-term threats to the world according to GRPS respondents, but these fall back in the rankings towards the long term and none appear among the most potentially severe, signalling a possible blind spot."

That sentence is the single most useful line in the whole seven-year corpus for judging the survey's own reliability. Two years before AI enters the list at #6 on the ten-year horizon, the WEF is on record that its expert panel is under-weighting technology at long range.

2023 — cyber promoted, AI absent

The GRPS 2022–2023 closed in autumn 2022 — before generative AI reached the general public. The consequence is visible everywhere in this edition.

Widespread cybercrime and cyber insecurity is a new entrant to both top 10s, at #8 on the two-year and #8 on the ten-year horizon. This is the first time a technological risk sits in the severity top 10 on both horizons.

Everything else technological sits low: Misinformation and disinformation #16 / #11, Breakdown of critical information infrastructure #24 / #16, Digital power concentration #29 / #17, Digital inequality #31 / #30, Adverse outcomes of frontier technologies #32 / #18.

The in-depth chapter, 2.5 Digital rights: privacy in peril, is about surveillance, biometrics and the commodification of personal data. AI is present only as an aggravating factor — "the proliferation of data-collecting devices and data-dependent AI technologies could open pathways to new forms of control over individual autonomy." There are 13 mentions of AI in the whole document.

2024 — the break

The GRPS 2023–2024 is the discontinuity. Three things happen at once.

1. Misinformation takes the #1 short-term slot. Misinformation and disinformation is ranked the single most severe global risk over two years — the first time in the period that any technological risk tops the ranking. The stated reason is electoral: "close to three billion people are expected to head to the electoral polls" across India, the US, the UK, Mexico, Indonesia, Pakistan and Bangladesh over the following two years. Average severity 4.7/7.

2. AI enters the risk list and goes straight into the long-term top 10. Adverse outcomes of AI technologies is a new risk, ranked #29 over two years but #6 over ten — average 10-year severity 5.3/7. The report is explicit that this is a new entrant to the top-10 rankings.

3. AI enters the current-crisis question in second place. Asked which risks are most likely to present a material crisis in 2024, respondents put "AI-generated misinformation and disinformation" at #2 (53%), behind extreme weather (66%) and ahead of societal polarization (46%). The compound naming is significant: AI arrives in the immediate-risk framing not on its own account but as an amplifier of the information risk.

The dedicated chapter, 2.4 AI in charge, sets out the concerns that will structure every subsequent edition: market concentration in a "singular, globally integrated supply chain that favors a few companies and countries"; the black-box problem (we do not understand the technology) compounded by the pacing problem (regulation cannot keep up); the observation that "the precautionary principle has largely not been applied in the development of AI, as regulators erred on the side of innovation"; and the integration of AI into "conflict decisions" with the risk of unintended escalation.

Two technological risks now sit in the two-year top 10 (misinformation #1, cyber insecurity #4) and three in the ten-year top 10 (misinformation #5, AI #6, cyber #8).

2025 — consolidation, and an explicit warning of complacency

The GRPS 2024–2025 largely holds the 2024 shape. Misinformation and disinformation is #1 over two years for the second year running and #5 over ten. Cyber espionage and warfare — the narrowed cyber risk — is #5 / #9. Adverse outcomes of AI technologies holds #6 over ten years but slips to #31 over two.

That short-term position produces the edition's most striking editorial intervention. Under the heading "Technological risks — still 'under the radar'", the report writes:

"In a year that has seen considerable experimentation by companies and individuals in making the best use of AI tools, concerns about Adverse outcomes of AI technologies are low in the risk ranking on a two-year outlook. However, complacency around the risks of such technologies should be avoided given the fast-paced nature of change in the field of AI and its increasing ubiquity."

This is the 2022 "blind spot" warning restated, now aimed specifically at AI. The WEF is declining to take its own survey at face value.

The in-depth chapter, 1.5 Technology and polarization, is about the interaction rather than the technology: 5.5 billion people online, rising computing power, "fast-improving AI and GenAI models" cutting the cost of producing content, and the resulting loop between misinformation, societal polarization and algorithmic bias. A new risk, Online harms, enters at #14 / #22. Censorship and surveillance sits at #16 / #14.

In the current-risk landscape for 2025, Adverse outcomes of AI technologies is #13.

2026 — AI becomes a present concern, and geoeconomics takes the top slot

The GRPS 2025–2026 shows two movements at once.

Misinformation loses #1 but gains at depth. Geoeconomic confrontation takes the two-year top slot (up eight positions), pushing Misinformation and disinformation to #2. But over ten years misinformation rises to #4, its highest long-term placement in the period. Cyber insecurity — the name reverts — is #6 / #8. Online harms rises to #12 / #18. Censorship and surveillance #14 / #15.

AI arrives in the present tense. In the current-risk landscape for 2026, Adverse outcomes of AI technologies is a new entrant to the top 10 at #8 (4% of respondents), alongside Cyber insecurity at #9 (3%). This is the first edition in which AI is named as a candidate for a material crisis this year by a measurable share of the panel.

In the severity ranking, AI holds the pattern but pushes further: #30 over two years, #5 over ten — average severity 3.50 versus 5.28 on the 1–7 scale. The report states plainly that this is "the risk with the largest rise in ranking over time" of all 33 risks surveyed.

The report's own summary of the technological picture:

"Technological risks are growing, largely unchecked … Technological developments and new innovations are driving opportunities, with vast potential benefits from health and education to agriculture and infrastructure, but also leading to new risks across domains, from labour markets to information integrity to autonomous weapons systems."

The dedicated chapter, 2.7 AI at large, is analysed in section 7 below. A companion chapter, 2.6 Quantum leaps, treats quantum acceleration as "another facet of strategic rivalry, economic bifurcation and political polarization" — yet Adverse outcomes of frontier technologies remains #33 / #25, the lowest-ranked technological risk in both horizons.

5. The comparable series, 2023–2026

Severity ranks, short term (2 years) / long term (10 years). Lower is more severe. Total risks ranked: 32 (2023), 34 (2024), 33 (2025), 33 (2026).

Technological risks in the two-year top 10: 1 (2023) → 2 (2024) → 2 (2025) → 2 (2026). Technological risks in the ten-year top 10: 1 (2023) → 3 (2024) → 3 (2025) → 3 (2026).

The series stabilises after the 2024 break. What moves inside it is composition, not volume: misinformation deepens on the long horizon, cyber holds a stable mid-single-digit band, AI is immovable near the top of the ten-year list and immovable near the bottom of the two-year one, and the "hard" frontier-technology risk stays where it has always been — last.

6. Four cross-cutting movements

6.1 From infrastructure to information

The centre of gravity shifts decisively. In 2020 the technological risks that scored were about systems failing: infrastructure breakdown, cyberattacks, data theft. By 2026 the risks that score are about truth failing: misinformation at #2 / #4, censorship and surveillance at #14 / #15, online harms at #12 / #18. Cyber is the one constant across the whole period, but it is no longer the story — it is the floor.

6.2 Information risk is now judged a societal risk in technological clothing

Every edition from 2024 onward describes misinformation through its societal effects rather than its technical mechanism: electoral legitimacy (2024), polarization and algorithmic bias (2025), "a fragmented public sphere in which consensus on basic facts breaks down" (2026). The 2024 report identifies Misinformation and disinformation and Societal polarization as "the most strongly connected risks in the network, with the largest potential to amplify each other." The WEF has classified the risk by its cause and analysed it by its effect.

The 2026 edition adds external corroboration rather than survey opinion: trust in news down from 44% to 40% between 2018 and 2025, news avoidance up from 32% to 40%, concern over misinformation up from 54% to 58% (Reuters Institute); in the US, social media as primary news source rising from 4% (2015) to 34% (2025); use of AI tools for finding information rising from 11% (2024) to 24% (2025).

6.3 The repression counter-risk hardens into its own line item

From 2024 the reports consistently pair the disinformation risk with the risk of the response to it. The 2024 formulation: "there is a risk of repression and erosion of rights as authorities seek to crack down on the proliferation of false information — as well as risks arising from inaction." Censorship and surveillance rises steadily across the comparable series (21 → 16 → 14 short term), the only technological risk with a monotonic short-term climb. The 2026 top-10 short-term list also contains Erosion of human rights and/or of civic freedoms at #8.

This is a genuine analytical advance in the corpus: the information risk is framed as a dilemma, not a threat with an obvious remedy.

6.4 Concentration is raised, then quietly reclassified

Digital power concentration (2021–2023) becomes Technological power concentration (2024, at #12 / #12 — its strongest showing), then vanishes. The 2025 and 2026 lists carry Concentration of strategic resources and technologies instead — filed as an economic risk and ranked #12 / #13 (2025) and #15 / #11 (2026).

The concern survives; the framing moves from technology governance to economic statecraft. This is coherent with 2026's overall theme (The Age of Competition, with geoeconomic confrontation at #1), but it has an analytical cost: the question of who controls AI infrastructure is now measured as a resource-competition problem between states, not as a technology-governance problem. The 2024 chapter's careful account of the generative-AI value chain — ASML at ~90% of lithography, TSMC at ~56% of foundry, >80% of assembly and test in Asia, >70% of top-cited AI research from the US — has no ranked risk carrying it forward.

7. Artificial intelligence: the arrival of a category

7.1 The trajectory in one table

7.2 Four observations

(a) AI did not climb into the ten-year top 10 — it arrived there. In its first edition as a named risk (2024) it enters at #6 and has never left the top six since. There is no gradual ascent to observe, because the instrument did not exist before 2024. The prehistory is real but undifferentiated: Adverse tech advances at #4 in the 2021 five-to-ten-year "existential threats" band is the same anxiety without the name.

(b) The three-year short-term position is remarkably stable and remarkably low. #29, #31, #30 out of 34, 33, 33. Whatever else has changed, the expert panel has not, in three successive years, judged AI a material two-year threat — while ranking it top-six at ten years every time. This is the central quantitative fact about AI in this corpus, and section 8 addresses it.

(c) The severity gap is now measured, and it is the largest in the survey. The 2026 edition publishes both scores for AI: 3.50 at two years, 5.28 at ten — a gap of 1.78 points on a seven-point scale, and the largest rank movement (+25 places) of any of the 33 risks. For comparison, Societal polarization moves 4.65 → 5.12 (+0.47).

(d) 2026 is the year AI becomes a present-tense concern, on a different instrument. The severity ranking still puts it at #30 over two years. But the current-risk question — a different exercise, asking what will produce a material crisis this year — puts it at #8, a new entrant, at 4% of respondents. And the Executive Opinion Survey shows three countries ranking it their single most important national risk and 20 placing it in their top five, rising to #4 in South-Eastern Asia. The aggregate expert panel and national business executives are diverging, and the executives are more worried, sooner.

7.3 How the content of the AI risk changed, 2024 → 2026

The 2024 chapter, AI in charge, is a governance and market-structure argument. Its three headline claims: market concentration and national-security incentives will constrain guardrails; AI will create "a new set of divides between those who are able to access or produce technology resources and intellectual property and those who cannot"; deeper integration of AI in conflict decisions could cause unintended escalation. The framing is institutional — who owns it, who regulates it, who is left out.

The 2026 chapter, AI at large, is a social-consequence argument, and the shift is the most substantive development in the AI material across the corpus. Its three headline claims:

• "In a worst-case scenario for labour markets, increases in both productivity and unemployment could drive permanently K-shaped economies."
• "The potential for creativity, learning and leisure could give way to loss of purpose, meaning and contribution to society, coupled with erosion of alignment around objective facts."
• "The rising range of military use cases for AI will come with commensurate risks, in the worst case leading to rapid and perhaps unintentional escalation of conflicts."

The report explains its own change of emphasis: "AI has shifted from a frontier technology to a systemic force shaping economies, societies and security" — with the global AI market projected from ~$280 billion (2024) to $3.5 trillion (2033).

Three strands are worth isolating.

Jobless productivity. The chapter sets a benign estimate (170 million roles created, 92 million displaced, net +78 million by 2030; 86% of companies expecting AI to transform their business models) against a severe one (up to 50% of US entry-level white-collar jobs eliminated within five years, unemployment at 10–20%). It coins the image of a "white-collar rust belt" in cities that are today knowledge and services hubs, "generating a powerful, angry, political force," and warns of "permanently K-shaped economies" in advanced countries. The transmission mechanism is named: Inequality is "the most interconnected risk for the second year in a row … labour displacement feeds inequality, which drives societal polarization."

Purpose in drift. This is new material with no precedent in the earlier editions — a risk register entry about meaning. The chapter treats occupational identity loss, the possibility of a "lost generation" subsisting on universal basic income, over-dependency on AI as a "second brain", AI companions increasing loneliness, and the erosion of critical thinking in education systems. It also carries the second-order physiological argument: generative models consuming "up to 4,600 times more energy than traditional software", a projected US public-health burden "over $20 billion annually by 2028", and water stress around data-centre build-outs.

Military misuse or mistakes. The most concrete escalation scenario in the corpus. Militaries embedding AI in "intelligence, surveillance, logistics, and command functions" shift the risk "from tactical to systemic"; data poisoning is named as a covert weapon; and the timing argument is explicit — "military crises that once unfolded over days or hours could instead escalate in seconds," with "a conflict started by technical error rather than strategic intent." The chapter records that countries are "starting to implement AI tools for managing nuclear weapons stockpiles and in some areas of nuclear weapons command, control, and communications." Its diagnosis of the incentive structure is unusually blunt: "major powers are rushing to integrate AI across military domains, each fearing strategic disadvantage if rivals move first. This dynamic incentivizes rapid deployment over rigorous safety testing … Each country's pursuit of security may, collectively, produce a more dangerous world."

Information chaos as the connective tissue. The chapter closes the loop with the misinformation risk: "realistic deepfakes and AI-generated Misinformation and disinformation are already flourishing; within a decade they could become ubiquitous, making it impossible for citizens to distinguish truth from deception … any scandal can be dismissed as a deepfake and any deepfake might be real." This is why the two risks cannot be analysed separately from 2024 onward, and why the WEF's interconnection maps place them adjacent in every edition since.

7.4 The one thing that has not moved

Adverse outcomes of frontier technologies — the quantum, biotech, geoengineering and brain-computer-interface risk — is ranked 32nd, 34th, 33rd, 33rd on the short horizon and 18th, 24th, 23rd, 25th on the long horizon across 2023–2026. It is the lowest-ranked technological risk in every comparable edition, in both horizons, without exception.

This holds even in editions that devote an in-depth chapter to it: 2025's Losing control of biotech? (10-year severity 4.5, rank #23) and 2026's Quantum leaps. The panel's long-horizon imagination is, on this evidence, specific to AI rather than generally receptive to frontier technology. That is worth stating, because it argues against the easy reading that AI's #5 placement is merely a fashion effect: if it were, quantum and biotech would be rising with it. They are not.

8. The short-term / long-term gap, and what it means

The defining feature of the AI risk in this corpus is the gap: #30 short term, #5 long term (2026); 3.50 versus 5.28 in severity. There are at least four readings, and the reports themselves hold more than one.

Reading 1 — the panel is right, and the risk is genuinely deferred. Labour-market displacement at scale, military integration and information ubiquity are decade-scale phenomena. On this reading the ranking is a correct statement about timing, not a failure of perception.

Reading 2 — the panel is systematically late, and says so. This is the WEF's own reading, and it has now made it twice: the 2022 "possible blind spot" on technological risk at long range, and the 2025 "complacency around the risks of such technologies should be avoided." The corpus contains a documented instance of the panel being late — State-based armed conflict "was overlooked as a leading two-year risk two years ago" (2025 edition) before becoming the #1 current risk at 23%. There is no reason to assume technological risk is immune to the same lag.

Reading 3 — the instrument has a short-horizon bias for diffuse risks. A two-year severity question rewards risks with identifiable events: an election, an invasion, a tariff, a storm. AI's short-term harms are distributed and attributed elsewhere — a laid-off worker is Lack of economic opportunity, a deepfake is Misinformation and disinformation, an AI-enabled intrusion is Cyber insecurity. The measured low short-term rank for AI may partly be an artefact of its harms being counted under other risks that themselves rank high. Both misinformation (#2) and cyber insecurity (#6) sit in the 2026 two-year top 10, and both are described in that same report as AI-amplified.

Reading 4 — the divergence between panels is the signal. The GRPS expert panel puts AI at #30 over two years. The Executive Opinion Survey, asking national business leaders about the same two-year window, has three countries ranking it first and twenty placing it top-five. The current-risk question puts it at #8 for 2026. Three instruments in one report give three answers, and the two closest to operating businesses give the more urgent ones.

For a risk-management reader the practical conclusion does not depend on choosing between these. Under readings 2, 3 and 4 the two-year rank understates exposure; under reading 1 it does not, but the ten-year rank still sits at #5 and the report's own chapter argues that "the long-term risks stemming from AI depend on choices made or avoided within the short to medium term." Either way, the decision-relevant horizon is nearer than #30 suggests.

9. What the WEF does not say

Stated plainly, so the analysis is not read as more than it is.

• The GRPS is an opinion survey, not a measurement. It records what a self-selected expert network believes, weighted equally across respondents. Shifts in ranking are shifts in perception. The reports are careful about this; secondary coverage frequently is not.
• No probabilities, and — before 2024 — no severity scores. The 1–7 severity averages appear only for selected risks in selected figures, and the earlier editions give none. Anyone wanting a quantified risk model will not find one here.
• Rank is zero-sum. Misinformation falling from #1 to #2 in 2026 does not mean misinformation eased; it means geoeconomic confrontation rose eight places. Its ten-year rank improved in the same edition, and its measured severity did not fall.
• The list bounds the finding. Digital inequality has not been ranked since 2023 because it is no longer on the list, not because it was resolved. The same applies to technological power concentration after 2024.
• The composition of the respondent pool is not held constant across editions, and the reports do not publish a comparability adjustment. Year-on-year rank changes carry this unquantified noise.

10. Findings in brief

1. The technological risk category was remade, not extended, between 2022 and 2024. Four of six risks were replaced, and the vocabulary moved from infrastructure and access to content, control and cognition.
2. 2022 is the trough and the tell. No technological risk in the ten-year top 10, and the WEF itself flags "a possible blind spot" — two years before AI enters at #6.
3. 2024 is the break. Misinformation takes #1 over two years; AI enters the list and lands at #6 over ten; "AI-generated misinformation and disinformation" is the #2 current crisis at 53%.
4. The post-2024 series is stable in volume and unstable in composition: consistently two technological risks in the two-year top 10 and three in the ten-year top 10, but with misinformation deepening long term, censorship and surveillance climbing short term, and concentration reclassified out of the category.
5. AI's signature is the horizon gap, and it is the largest in the 2026 survey: #30 versus #5, severity 3.50 versus 5.28. It has been top-six on the ten-year horizon in every edition in which it has existed.
6. 2026 is the first year AI is a present-tense concern — a new entrant at #8 in the current-risk landscape, and top-five national risk in twenty economies per the Executive Opinion Survey.
7. The AI argument itself changed from a governance and market-structure case (2024) to a social-consequence case (2026): K-shaped labour markets, a "white-collar rust belt", loss of purpose, and military escalation "started by technical error rather than strategic intent."
8. Information integrity is the through-line. It is the only technology theme in the two-year top 10 in every edition since 2024, and it is the mechanism by which the AI risk is expected to reach societies first.
9. Frontier technology other than AI has not moved at all — last or near-last among technological risks in both horizons for four consecutive editions, including in the years it received dedicated chapters. The panel's long-horizon concern is specific to AI.
10. Three instruments inside the 2026 report disagree about AI's urgency, and the two closest to operating businesses are the more alarmed. That disagreement is more informative than any single rank.

11. Source table

Companion files, same folder

• WEF-GRR-2020-2026-tech-info-AI-risk-rankings.xlsx — the extracted rank data, six sheets, with the methodological warning carried on the README sheet.
• WEF-GRR-2020-2026-AI-tech-risks-synthesis.html — a self-contained synthesis page.

All quotations are drawn from the seven reports listed above and are reproduced for analysis and comment. The Global Risks Report is © World Economic Forum; rights are reserved to the Forum and no part of those reports may be redistributed without its permission.