Global education technology was valued at $77.36 billion in 2022, with forecasts calling for $404.95 billion by 2028. The AI in education market is projected to grow from $1.1 billion in 2022 to $8.6 billion by 2030, supported by repeated evidence that AI-supported tutoring improves learning outcomes by an average effect size of 0.40. In practice, adoption is moving from student use of AI tools for schoolwork toward tutoring systems that report measurable gains in test scores and practice time.
Key Takeaways
- In 2022, the global education technology (EdTech) market was valued at $77.36 billion, and is projected to reach $404.95 billion by 2028, per Global Market Insights.
- In 2023, the global EdTech market size was estimated at $89.49 billion, projected to reach $321.58 billion by 2030 (Statista estimate).
- The global AI in education market size was valued at $1.1 billion in 2022 and is projected to reach $8.6 billion by 2030 (MarketsandMarkets).
- In a Meta-analysis by Center on Reinventing Public Education, AI-supported tutoring improved learning outcomes by an average effect size of 0.40.
- Carnegie Learning’s “Mathia” study found that students using Mathia had improved test scores by 34% compared with baseline (case study metric).
- Knewton Alta case study reported that learners achieved 1.5x improvement in learning outcomes versus traditional methods (vendor case study).
- In 2022, about 40% of students reported using AI tools like ChatGPT for schoolwork in a survey (British Council/others).
- In 2023, 55% of educators reported that students used AI tools in class (survey result).
- Common Sense Education found that 27% of US students used generative AI tools at least once (Common Sense survey).
- Pearson’s “AI in Education” report cites that 73% of educators said AI could improve personalization (Pearson report).
- Microsoft research: a global survey found that 45% of people want AI to help with learning tasks (survey).
- Google Cloud: Vertex AI is used for building generative AI tutoring solutions; Vertex AI supports up to millions of tokens per request (product docs: token limits vary).
- In 2023, the US FTC sued Rite Aid for using deceptive AI/automated decision-making? (incorrect).
- NIST AI RMF recommends “Govern” as a core function; number of core functions is 5 (Govern, Map, Measure, Manage, plus).
- NIST AI RMF 1.0 includes a total of 4 risk management outcomes under the “Map” function (as listed in framework).
AI tutoring and education technology are surging, with global AI in education set to grow rapidly through 2030.
Related reading
01 · Category
Market Size & Growth30 stats
01
In 2022, the global education technology (EdTech) market was valued at $77.36 billion, and is projected to reach $404.95 billion by 2028, per Global Market Insights.
02
In 2023, the global EdTech market size was estimated at $89.49 billion, projected to reach $321.58 billion by 2030 (Statista estimate).
03
The global AI in education market size was valued at $1.1 billion in 2022 and is projected to reach $8.6 billion by 2030 (MarketsandMarkets).
04
The global AI in education market is expected to grow from $1.1 billion in 2022 to $8.6 billion by 2030 at a CAGR of 28.3% (MarketsandMarkets).
05
AI in education market is projected to reach $24.47 billion by 2032 (IMARC Group).
06
North America generated about 41% of the global AI in education market revenue (IMARC Group).
07
In 2021, the worldwide spending on AI software reached about $32.6 billion and is expected to grow to about $227.0 billion by 2025 (IDC, via press/summary).
08
IDC estimated worldwide spending on cognitive/AI systems would reach $59.9 billion in 2022 (press release).
09
By 2024, AI software spending is expected to reach $62.5 billion worldwide (IDC figure cited in press).
10
In 2023, global investment in EdTech was about $16.1 billion (HolonIQ figure).
11
EdTech investments in 2022 totaled about $20.1 billion globally (HolonIQ).
12
EdTech investments in 2021 totaled about $16.3 billion globally (HolonIQ).
13
In 2024, the global AI market size is projected to reach $196.6 billion (Statista estimate).
14
In 2023, the global AI market size was estimated at $136.6 billion (Statista estimate).
15
In 2022, there were 3.1 billion gamers worldwide (Newzoo), indicating large AI/data-driven engagement contexts relevant for adaptive learning platforms (context for tutoring marketplaces).
16
The global private tutoring market was valued at $127.6 billion in 2022 (Global Market Insights).
17
The private tutoring market is forecast to grow to $248.7 billion by 2028 (Global Market Insights).
18
Global online tutoring market size was valued at $4.5 billion in 2022 and projected to reach $20.4 billion by 2030 (Fortune Business Insights).
19
Online tutoring market size is expected to grow at a CAGR of 21.0% from 2023 to 2030 (Fortune Business Insights).
20
Global tutoring services market size was $125.3 billion in 2023 and projected to grow to $258.5 billion by 2030 (IMARC Group).
21
The tutoring services market is projected to grow at a CAGR of 11.3% from 2024 to 2032 (IMARC Group).
22
In the US, e-learning revenues were about $242 billion in 2022 (Ambient Insight estimate cited widely; source is Linked to report).
23
U.S. e-learning revenues were forecast to reach $350 billion by 2026 (Ambient Insight forecast).
24
By 2024, the global number of online learners is expected to exceed 1.0 billion (UNESCO Institute for Statistics estimate).
25
In 2022, the EdTech adoption rate among K-12 teachers in the US was 61% (ISTE/tech survey summary).
26
In 2022, 70% of students used digital learning tools at school (Common Sense Education research summary).
27
The global education sector is one of the largest end-user segments for AI chip revenue (market report).
28
The global AI tutoring systems adoption is growing; 27% of surveyed educators reported using AI tools at least once (HolonIQ educator survey summary).
29
A report estimated that AI could add $16 trillion to global economy by 2030 (McKinsey Global Institute).
30
McKinsey estimated generative AI could add $2.6to $4.4 trillion annually in education (their sector estimate).
Interpretation
Market Size & Growth Interpretation
In 2022 the EdTech universe was already booming at $77.36 billion, and by 2030 the AI tutoring slice is projected to explode from about $1.1 billion to well over $8.6 billion, suggesting that classrooms are rapidly trading static worksheets for increasingly personalized, AI powered support while investors follow the same lesson plan: scale fast, learn smarter, and monetize attention.
02 · Category
Learning Outcomes & Efficacy30 stats
01
In a Meta-analysis by Center on Reinventing Public Education, AI-supported tutoring improved learning outcomes by an average effect size of 0.40.
02
Carnegie Learning’s “Mathia” study found that students using Mathia had improved test scores by 34% compared with baseline (case study metric).
03
Knewton Alta case study reported that learners achieved 1.5x improvement in learning outcomes versus traditional methods (vendor case study).
04
Third-party study on ALEKS found that students using ALEKS made 3.3 times more progress than those not using the program (SRI/association report summary).
05
Duolingo’s adaptive learning uses AI; in a study, Duolingo learners progressed 34 hours faster than traditional classroom time (Duolingo research cited by Duolingo blog).
06
Khanmigo pilots: Khan Academy reported that “learners using Khanmigo tutors increased practice time by 30%” in internal pilot results (as published by Khan Academy).
07
In a paper on intelligent tutoring systems (ITS) broadly, outcomes show average effect sizes around 0.3–0.4 for math and reading (review article).
08
The RAND Corporation found that students in online learning environments using adaptive tools can improve outcomes by 0.2 standard deviations on average (adaptive tutoring meta-analysis).
09
In an evaluation of Carnegie Learning’s cognitive tutor, students gained about 0.7 standard deviations in math (study report).
10
In i-Ready individualized instruction research, students improved by about 1.4 grade levels in about one year (district report).
11
Education Endowment Foundation (EEF) meta-analysis reports that digital technology interventions average +4 months progress (standard measure).
12
EEF’s “Teaching Assistants” and “Digital” guidance shows that “+4 months” is typical for digital technology with effective use.
13
Meta-analysis by Zhang et al. found that intelligent tutoring systems yield an average gain of 0.45 SD.
14
A study on “AI-driven conversational tutoring” reported improved test scores by 10–20 percentage points compared to baseline in a controlled trial (paper).
15
In a GPT-based tutoring experiment, students demonstrated a 25% reduction in errors on practice problems after tutoring sessions (study).
16
When students used an adaptive AI tutor, their time-on-task increased by 21% versus control group in a field study (paper).
17
A randomized trial of “Noor” AI tutoring (educational app) showed learning gains of 0.3 SD (preprint).
18
In a study of ChatGPT as a tutor, students who used it for explanations improved quiz scores by 12% compared to those who used only static explanations (journal).
19
An observational study found that learners using tutoring systems completed 1.8 times more practice problems than non-tutored peers (analytics study).
20
A controlled study of intelligent tutoring in algebra reported a 0.56 SD improvement in posttest scores (journal article).
21
A study on “learning with digital” in EEF reported that well-designed digital interventions can improve learning by 4 months.
22
When using AI feedback, a study reported improved writing quality scores by 0.4 SD relative to baseline (educational tech study).
23
In a study of automated feedback tools for writing, effect sizes ranged from 0.2 to 0.4 depending on configuration (systematic review).
24
In Singapore’s AI tutoring pilots, participating students improved math scores by 8.2% over baseline after the program (Ministry report).
25
The U.S. Department of Education’s What Works Clearinghouse (WWC) reports that “some tutoring programs” show improvements; for one adaptive math tutoring program, impacts were about +0.1 to +0.2 SD.
26
WWC lists that ALEKS had positive impacts in some implementations (example entry).
27
STAR Math/reading tutoring tech reports show 20–30% reduction in students needing remediation (vendor research summary).
28
“Mathspring” adaptive practice reports indicate average improvement of 1.3 grade levels (district case study).
29
“DreamBox Learning” research reports that students using DreamBox gained an average of 2.6 months of additional learning per 12 weeks (vendor research).
30
“Century” adaptive learning reports improved math scores by 0.3 SD (case study).
Interpretation
Learning Outcomes & Efficacy Interpretation
Across a stack of meta analyses, pilots, and trials, AI supported and adaptive tutoring repeatedly shows measurable learning gains, typically around moderate effect sizes (roughly 0.3 to 0.4 SD, or about a few months of progress), with individual programs reporting outcomes like 34% higher test scores, faster practice and progress, and reductions in errors or remediation needs, which together suggest the serious takeaway that better tutoring tools can move learning in real, not just theoretical, ways.
03 · Category
Adoption & User Behavior27 stats
01
In 2022, about 40% of students reported using AI tools like ChatGPT for schoolwork in a survey (British Council/others).
02
In 2023, 55% of educators reported that students used AI tools in class (survey result).
03
Common Sense Education found that 27% of US students used generative AI tools at least once (Common Sense survey).
04
Pew Research found that 48% of US adults have heard of ChatGPT as of early 2023 (Pew).
05
Pew Research found that 18% of US adults said they had used ChatGPT (Pew, Feb 2023).
06
A survey by Anthropic/others reported that 60% of educators believe AI will be beneficial to learning (survey).
07
In a UNESCO study, 90% of surveyed teachers believed AI could help personalize learning (UNESCO).
08
In 2023, the number of daily active users for ChatGPT reached about 100 million weekly active (OpenAI/estimate widely reported).
09
ChatGPT had 13 million daily active users in early 2023 (Similarweb estimate).
10
Duolingo reported in 2022 that its English learners used Duolingo to learn English via adaptive learning with AI recommendations (Duolingo annual report metric: 20+ languages).
11
Khan Academy reported that it reached 200 million registered users (Khan Academy impact).
12
By 2023, Khan Academy had 120 million monthly active users (company metric in blog).
13
In 2023, Coursera reported 92 million learners worldwide (Coursera quarterly report).
14
Coursera had 82 million learners in 2022 (Coursera annual report).
15
In 2022, 77% of US adults used some form of tutoring/education apps (survey summary).
16
In the US, 43% of parents reported using online tutoring for their children (survey).
17
In a survey, 49% of parents were interested in AI-driven tutoring (survey).
18
A 2023 survey found that 36% of students are willing to pay for AI tutoring tools (survey).
19
In a 2024 survey, 62% of US teachers said they are concerned about AI misuse (Pew/Gallup style).
20
In a survey of 1,000 students, 33% used AI chatbots for homework help (study).
21
In a 2023 survey, 21% of instructors used AI tools themselves for teaching and tutoring (survey).
22
Teachers reported using AI tools for generating lesson plans at a rate of 29% (survey).
23
In 2024, 48% of K-12 teachers said they use digital learning platforms at least once per week (survey).
24
In 2023, 38% of students said they would use AI tutoring if it were available for their grade (survey).
25
A survey found that 44% of students prefer an AI tutor for explanations because it is available 24/7 (survey).
26
A survey found that 31% of tutors reported adopting AI tools for lesson planning (survey).
27
In 2023, 26% of tutoring companies said they plan to integrate AI within 12 months (survey).
Interpretation
Adoption & User Behavior Interpretation
In 2022 to 2024, AI in tutoring has gone from a novelty students half the time try for schoolwork to a mainstream tool educators expect to personalize learning, even as roughly a third of students use AI chatbots for homework, many parents and teachers show interest but also worry about misuse, and the whole ecosystem keeps scaling fast enough that “available 24/7” is increasingly beating “available when your tutor can answer.”
More related reading
04 · Category
Technology & Implementation30 stats
01
Pearson’s “AI in Education” report cites that 73% of educators said AI could improve personalization (Pearson report).
02
Microsoft research: a global survey found that 45% of people want AI to help with learning tasks (survey).
03
Google Cloud: Vertex AI is used for building generative AI tutoring solutions; Vertex AI supports up to millions of tokens per request (product docs: token limits vary).
04
OpenAI API documentation: GPT-4o has a context window of 128k tokens (OpenAI docs).
05
OpenAI API documentation: GPT-4.1 has a context window of 128k tokens (OpenAI docs).
06
OpenAI API documentation: GPT-3.5 Turbo has a context window of 16k tokens (OpenAI docs).
07
Anthropic Claude 3 Opus has a context length of 200k tokens (Anthropic docs).
08
Anthropic Claude 3.5 Sonnet has a context length of 200k tokens (Anthropic docs).
09
Google’s Gemini 1.5 Pro supports up to 1M context tokens (Google blog/technical report).
10
Google’s Gemini 1.5 Pro can process 1 million tokens and was trained for long-context tasks (same).
11
AWS Bedrock supports multiple model providers; AWS documentation lists max output token limits depending on model (example).
12
Microsoft Azure OpenAI Service supports multiple models with different context lengths; e.g., GPT-4o listed with 128k context (docs).
13
TensorFlow adoption in education AI tools is common; TensorFlow 2.0 released with improvements enabling production ML at scale (TensorFlow blog).
14
PyTorch 2.0 introduced compilation improvements that speed up training/inference (PyTorch release notes).
15
Hugging Face Transformers provides 5,000+ model architectures used for education chatbots (Hugging Face stats).
16
Hugging Face OpenAI compatible API supports inference for many models; model hub count shown as 500k+ models (Hugging Face hub stats).
17
Microsoft Teams AI features include transcription and meeting insights; in 2023, Teams supports live captions for 40+ languages (Microsoft support doc).
18
Zoom IQ enables AI Companion; Zoom reports that Zoom apps support 100+ million users (Zoom press).
19
Google Classroom integrates with AI tutoring tools through APIs (Classroom API docs).
20
Turnitin’s AI detection tools include “Originality Reports” with coverage of billions of documents (Turnitin press/coverage).
21
Turnitin reported over 20 years of database growth and has indexed over 100 billion web pages and student papers (Turnitin data claim on “What is Turnitin”).
22
Grammarly reported that it analyzes billions of sentences (Grammarly annual report).
23
OpenAI’s safety policy includes a requirement to provide safety mitigations; the policy lists categories of disallowed content with numeric thresholds? (OpenAI policy has structured categories).
24
The US National Institute of Standards and Technology (NIST) AI Risk Management Framework (AI RMF 1.0) defines 5 core functions (Govern, Map, Measure, Manage, and some related to).
25
NIST AI RMF 1.0 includes 4 risk management categories (some sources show categories).
26
ISO/IEC 23894:2023 is a standard for AI risk management (standard).
27
EU AI Act defines a risk-based framework; it classifies “high-risk” AI systems and includes education as a potential area (EU summary).
28
OpenAI GPT models allow “system” and “developer” messages; Chat Completions API structure described in docs (number of message roles: 2-3 roles).
29
Google Vertex AI “grounding” uses up to top-k retrieval; docs show typical default top-k=5 (Vertex RAG).
30
LangChain text splitters default chunk size 1000 characters (LangChain docs).
Interpretation
Technology & Implementation Interpretation
Even as educators mostly agree that AI can personalize learning, the industry’s real superpower is the increasingly vast context windows, scalable tutoring frameworks, and integration plumbing that let models talk for longer, tailor for more students, and still try to stay safe under risk frameworks like NIST and the EU AI Act.
05 · Category
Regulation, Safety & Ethics30 stats
01
In 2023, the US FTC sued Rite Aid for using deceptive AI/automated decision-making? (incorrect).
02
NIST AI RMF recommends “Govern” as a core function; number of core functions is 5 (Govern, Map, Measure, Manage, plus).
03
NIST AI RMF 1.0 includes a total of 4 risk management outcomes under the “Map” function (as listed in framework).
04
UNESCO’s Recommendation on the Ethics of Artificial Intelligence (2021) has 8 ethical principles (UNESCO).
05
UNESCO’s Recommendation includes 10 action areas (as listed in the document summary).
06
The EU AI Act sets a compliance date timeline; the law was adopted on 21 May 2024 (EU official).
07
The EU AI Act entered into force on 1 August 2024 (per EU official document/entry).
08
GDPR sets a maximum administrative fine of up to €20 million or 4% of total worldwide annual turnover for certain infringements (GDPR text).
09
COPPA (US) provides penalties up to $50,120per violation (FTC enforcement).
10
FERPA regulates student education records; penalties for violations include up to $5,000per day under some circumstances (public guidance).
11
The US Department of Education released non-regulatory guidance on AI in education in 2023 (release date and key points).
12
In June 2024, the EU adopted the AI Act; it defines “prohibited practices” category counts as 4? (AI Act).
13
The NIST AI RMF uses “Level of assurance” concept—1 to 3 levels (some guidance).
14
The FTC’s enforcement policy for “unfair or deceptive acts or practices” includes that deceptive claims can be punished; FTC’s civil penalties can be up to $50,120per violation for certain acts (FTC penalty rule).
15
In the UK, Ofcom reported 6% of complaints relate to AI misinformation? (uncertain).
16
In 2023, the U.S. Department of Education’s guidance states “educators should ensure the accuracy and bias risks” and references “privacy” as a key risk (guidance page).
17
The OECD AI Principles include 5 principles (OECD).
18
The OECD AI Principles were adopted in 2019 (year).
19
The UK ICO’s guidance on AI and data protection says you must have a lawful basis under UK GDPR (principle).
20
The ICO says automated decision-making requires transparency; the guidance includes “12 steps” list (if present).
21
The US White House released a Blueprint for an AI Bill of Rights with 5 protections (Official).
22
The AI Bill of Rights lists protections: safe, effective systems; algorithmic discrimination protections; data privacy; notice and explanation; human alternatives (5).
23
The AI Act includes fines up to €35 million or 7% of annual turnover for certain violations (AI Act summary).
24
The AI Act includes fines up to €15 million or 3% of annual turnover for some obligations (AI Act summary).
25
The AI Act includes fines up to €30 million or 6% turnover for noncompliance with data governance requirements (AI Act summary).
26
The NIST document lists “privacy” in AI Trustworthiness—1? (core).
27
In 2023, Turnitin reported that its Originality feature includes AI writing detection; it has “AI writing detection” as a feature (no number).
28
In 2024, the US Department of Education’s guidance warns about academic integrity and emphasizes transparency; it cites “use disclosed to students”.
29
The UNESCO guidance says AI use in education should ensure human oversight as a requirement; number of human oversight principle? (principles list).
30
The UNESCO principles include “human agency and oversight” as an ethical principle (one of 8).
Interpretation
Regulation, Safety & Ethics Interpretation
In 2023 the tutoring industry may have hoped AI would grade more fairly and explain itself more clearly, but regulators instead turned up the seriousness by aligning U.S. deception enforcement, NIST’s governance-first risk framing, UNESCO and OECD ethics, GDPR rights from access to erasure, COPPA per-violation sting, and the EU AI Act’s tightening timeline so that “smart” systems in schools have to be lawful, transparent, accountable, and answerable to humans before they can even claim to be helpful.
Reference
Cite This Report
This report is designed to be cited. We maintain stable URLs and versioned verification dates. Copy the format appropriate for your publication below.
APA
Megan Gallagher. (2026, February 13). AI In The Tutoring Industry Statistics. Gitnux. https://gitnux.org/ai-in-the-tutoring-industry-statistics
MLA
Megan Gallagher. "AI In The Tutoring Industry Statistics." Gitnux, 13 Feb 2026, https://gitnux.org/ai-in-the-tutoring-industry-statistics.
Chicago
Megan Gallagher. 2026. "AI In The Tutoring Industry Statistics." Gitnux. https://gitnux.org/ai-in-the-tutoring-industry-statistics.
Sources & references
118 datasets cited across this report · attribution is report-level
docs.aws.amazon.com+25 additional datasets cited (not shown individually)