AI Literacy Report 2026:
What 1,017 Real Assessments Reveal
Key Findings: AI Literacy in 2026
- •Only 49% of professionals are AI Literate (score 28–59) — the largest group. They use AI but carry significant knowledge gaps.
- •44% cannot explain how AI works at a functional level (AI Fundamentals score below 5/10).
- •36% have no functional AI safety practice — unaware of data boundaries or downstream risks.
- •36% know AI's limitations only from headlines, not from personal experience.
- •Product managers score highest among major roles (59.7). Students score lowest (37.3).
- •Marketing & Content professionals have the lowest AI safety score (37.2) of any role.
- •The biggest gap between AI Literate and AI Native professionals is Technical Understanding — a 46-point divide.
- •Only 11% of professionals qualify as AI Native (score 80+).
Source: AISA AI Literacy Report 2026 — 1,017 assessments, Feb–Jun 2026. aisa.to/state-of-ai-literacy. Free to cite with attribution.
The AI Literacy Ladder: Where Professionals Stand
Nearly half of all professionals (49%) are AI Literate — they know enough to use AI, but not enough to use it safely, explain how it works, or recover when it fails.
The AI Literacy Ladder places every professional on a four-level spectrum. Bystanders are not yet engaging with AI. The AI Literate — the largest group by far — have crossed the adoption threshold but carry significant knowledge gaps. AI Fluent professionals understand both the how and the why. AI Natives have made AI integral to how they think, not just how they work.
The Literacy-to-Fluency jump is the hardest. Moving from Bystander to Literate requires curiosity. Moving from Literate to Fluent requires deliberate study of how AI systems work, where they fail, and when to trust their output. 62% of professionals are stuck below that threshold.
AI Literacy by Role: Who Understands AI Best?
Product managers lead among major roles (59.7). Students trail by 22 points (37.3) — the future workforce is not yet AI literate.
Across 1,017 assessed professionals, AI literacy varies sharply by role. Product managers — who bridge strategy, user needs, and technical constraints — score highest among roles with meaningful sample sizes. Engineers sit above average but carry a notable safety gap (43.8 vs 55.0 overall). Students, the professionals of tomorrow, score lowest across every dimension.
Average AI literacy score by role category (categories with n ≥ 15). Source: AISA, 1,017 assessments, 2026.
The Knowledge Deficit: Using AI Without Understanding It
AI Fundamentals is the weakest-scoring skill of all 11 measured (5.1/10). 29% of professionals who use AI in their workflow cannot explain how it works.
Across 1,017 measured assessments, a consistent pattern emerges: skills tied to doing things with AI (task decomposition, domain application, workflow integration) score higher than skills tied to understanding AI (how models work, what tools exist, when to apply which). Professionals have learned the buttons but not the machine behind them. When AI fails in unfamiliar ways, they have no mental model to fall back on.
Average score per criterion (1–10 scale). Source: AISA, 1,017 assessments, 2026.
How AI Fundamentals breaks down
44% of professionals score Novice or Developing on AI Fundamentals — they use AI vocabulary (tokens, models, training) without applying what those concepts mean in practice. Only 0.6% reach Expert level.
The AI Safety Literacy Gap
36% of professionals have no functional AI safety practice — unaware of data boundaries, unable to adjust scrutiny to stakes, not thinking about downstream impact.
Safety & Responsibility (44.4/100) is the weakest dimension of AI literacy. This is not about reading policy documents — it’s about knowing when to double-check, what not to share with a chatbot, and how an AI error in one place can cascade into real harm downstream. Organisations adopting AI at speed are building on a workforce that, by the numbers, does not yet understand the risks.
Average dimension scores (0–100). Source: AISA, 1,017 assessments, 2026.
The EU AI Act angle. Article 4 of the EU AI Act — already in force — requires employers to ensure “sufficient AI literacy” for all staff interacting with AI systems. Enforcement begins August 2026. With 36% of professionals below a functional safety threshold, the AI literacy skills gap is not abstract — it is a measurable compliance risk. An AI literacy assessment like AISA provides the evidence base that AI literacy in the workplace is being addressed.
The Limitation Blindspot: Knowing AI Fails — But Not From Experience
36% of professionals know AI’s limitations only from headlines — hallucination, bias — not from anything they have tested, experienced, or built defences against.
Limitation Awareness (T2) distinguishes between people who have heard that AI can be wrong and people who know from experience how and when it goes wrong. Only 17% of professionals score Proficient or above — meaning they can predict failure before it happens and adjust their approach preemptively. The rest are reactive at best.
A separate survey by EY found that only 28% of adults know AI can fabricate facts entirely. Stack Overflow’s 2025 developer survey found that only 29% of developers trust AI output — down 11 points year-over-year. The awareness problem extends across demographics and experience levels.
What Separates the AI Literate from the AI Native
The biggest gap between AI Literate and AI Native professionals is Technical Understanding — a 46-point divide. Understanding is the dimension that separates tiers, not usage.
Compare the dimension profiles of professionals in the Literate band (score 28–59) against those in the Native band (80+). The gap is not uniform — it is widest on the understanding dimensions (Technical, Safety) and narrowest on the application dimension (Workflow). AI Natives don’t just do more with AI. They understand more about AI. That understanding is what makes the difference.
Dimension averages by AI Literacy Ladder level.
The Growth Map: Where Professionals Need to Improve Most
58% of professionals identify Technical Understanding or Safety as their biggest growth area. The literacy dimensions are where the need is greatest.
After each assessment, AISA identifies the dimension with the most room to grow. Across the full population, the pattern is clear: understanding-related dimensions dominate. Technical Understanding is the #1 growth area (32%), followed by Safety (26%). Workflow — the doing-things-with-AI dimension — is the growth area for only 4% of professionals. The bottleneck is not adoption. It is comprehension. Closing this AI literacy skills gap requires targeted learning — starting with the foundational AI concepts that most professionals have skipped.
Percentage of professionals whose biggest growth area is each dimension. Source: AISA, 1,017 assessments, 2026.
The AI Literacy Score Distribution
The average professional scores 51.7 out of 100 — squarely in the AI Literate band. The distribution centres just above 50, with a long tail toward the Bystander end and a short tail toward AI Native. Because this sample is self-selected (professionals curious enough to measure themselves), the true workforce average is likely lower.
Methodology: Measured, Not Self-Reported
Every data point in this report comes from a completed conversational assessment — an average of 18 exchanges in which candidates demonstrate, not declare, how they understand and work with AI. An independent evaluator scores 11 criteria across 5 dimensions against the published AISA Rubric, and every assessment passes a two-stage scoring process with a final holistic calibration. The rubric independently covers 93% of the fluency markers identified in Anthropic’s AI Fluency Index (2026), plus four dimensions their methodology could not measure.
This distinction matters. The major AI literacy reports — DataCamp (59% skills gap), Deloitte (talent readiness at 20%), Bright Horizons/Harris Poll (only 17% use AI frequently) — all rely on self-reported surveys. A January 2026 academic study (LAK/arXiv) found that self-reported AI literacy measures show “low correlation” with objective measures, and that people who rated themselves highest actually scored lower on objective tests. This report sits on the measurement side of that gap: criterion-level granularity across 11 skills (not a single score), the separation of AI understanding from AI usage, per-dimension safety literacy measurement, and a validated assessment framework. To our knowledge, this is the largest objective AI literacy dataset published.
Sample: 1,017 assessments, February–June 2026, self-selected professionals across roles and industries. Detailed criterion-level analysis uses the 634 fully calibrated subset. Self-selection means this sample skews AI-curious — the true workforce average is likely lower than reported here. See also: The State of AI Fluency 2026.
AI Literacy Statistics 2026: The Key Numbers
Every statistic below comes from measured assessment scores. Free to cite with attribution.
- ▸49% of professionals are AI Literate — they have crossed the adoption threshold but carry significant gaps in understanding how AI works, when it fails, and how to use it safely.
- ▸44% of professionals cannot explain AI fundamentals — they use the vocabulary (tokens, models, training) without applying what those concepts mean in practice.
- ▸29% of professionals who actively use AI in their workflow score below functional on AI Fundamentals — they press the buttons without understanding the machine.
- ▸36% of professionals have no functional AI safety practice — unaware of data boundaries, unable to calibrate scrutiny to stakes, not considering downstream impact.
- ▸36% know AI’s limitations only from headlines — not from anything they have personally experienced, tested, or built safeguards against.
- ▸58% of professionals identify Technical Understanding or Safety as their biggest growth area — the understanding dimensions, not the application ones.
- ▸The gap between AI Literate and AI Native is widest on Technical Understanding (46 points) and narrowest on Workflow (38 points). Understanding, not usage, is what separates the levels.
- ▸Only 11% of professionals qualify as AI Native — making AI integral to how they think, not just how they work. The EU AI Act requires “sufficient AI literacy” for all staff touching AI systems. Enforcement begins August 2026.
Per-criterion percentages are from 1,017 assessments (calibrated subset, n = 493–634 per criterion).
How to cite this report
AISA (2026). The State of AI Literacy 2026. Retrieved from https://aisa.to/state-of-ai-literacy. Statistics and charts may be reproduced with a link to this page. For interviews or custom data cuts, get in touch.
AI Literacy Benchmark — FAQ
What is AI literacy?
AI literacy is the ability to understand how AI systems work, recognise their limitations, evaluate their outputs critically, and use them responsibly. It goes beyond knowing which button to press — it means understanding what happens when you press it, and what can go wrong.
What is the difference between AI literacy and AI fluency?
AI literacy is the foundation: understanding how AI works, its limits, and its risks. AI fluency builds on literacy — it adds the ability to apply AI effectively in your workflow, integrate it into complex tasks, and adapt your approach to different AI tools. You can be AI literate without being AI fluent, but you cannot be AI fluent without being AI literate.
How is AI literacy measured in this report?
AISA measures AI literacy through a 25-minute conversational assessment where professionals demonstrate — not self-report — their understanding of AI. An independent AI evaluator scores 11 criteria across 5 dimensions, including AI Fundamentals, Limitation Awareness, Output Evaluation, and Safety. Every score is tied to a direct quote from the conversation.
What is the AI Literacy Ladder?
The AI Literacy Ladder is a four-level model: AI Bystander (score 0–27, not engaging with AI), AI Literate (28–59, understands basics but gaps in knowledge), AI Fluent (60–79, competent practitioner), and AI Native (80–100, AI is integral to how they think and work). Nearly half of professionals (49%) sit at the AI Literate level.
Why does AI literacy matter for organisations?
AI adoption without AI literacy creates hidden risk. When employees use AI tools they do not understand, they cannot detect failures, assess data privacy implications, or judge when AI output needs human verification. 36% of professionals have no functional AI safety practice — that is an organisational risk, not just a skills gap.
Related Reading
Are you AI literate?
Find your level on the AI Literacy Ladder, get your score across 11 skills, and discover exactly where to focus next — in one 25-minute conversation.
See also: The State of AI Fluency 2026 — the companion report on operational AI skills.
The Science Behind AISA
In 2026, Anthropic published the AI Fluency Index — the largest empirical study of AI fluency to date, analysing 9,830 conversations. AISA covers 93% of the behaviours Anthropic identified as markers of AI fluency and goes even deeper with 4 additional dimensions. The U.S. Department of Labor's AI Literacy Framework (TEN 07-25) defines what every worker needs to know about AI — AISA covers 100% of its 25 sub-competencies.Read our analysis: Anthropic's AI Fluency Study & AISA · DOL AI Literacy Framework & AISA
AISA's framework is developed by a team with deep roots in tech, behavioural science, and AI product leadership — the rubric is informed by backgrounds spanning the Metropolitan Police, Harvard, Crowdbotics (Silicon Valley), and the European School of Economics.
