How many teens actually use AI for homework?

A lot, and the number is rising fast. Pew found teen ChatGPT use for schoolwork doubled from 13% (2023) to 26% (2024), and by fall 2025 just over half of teens had used a chatbot for schoolwork; RAND found 62% of students using AI for homework in late 2025.

Is using ChatGPT for homework cheating?

It depends on the assignment's rules and on whether AI does the thinking the task was meant to build. Teens themselves tend to see research help as fine and AI-written essays as not; the clearest fix is a teacher saying exactly how much AI help each assignment allows.

Has AI made kids cheat more?

Not clearly. Stanford surveys found 60-70% of high schoolers admitted some cheating long before ChatGPT and the rate did not jump in 2023, though many teens now believe classmates use AI to cheat often. Cheating seems driven more by pressure and assignment design than by any single tool.

Can teachers tell if my kid used AI? Are AI detectors accurate?

Not reliably. Independent tests find detectors inaccurate and easy to fool, they flag non-native English writers far more often, OpenAI withdrew its own detector, and people's gut judgments are no better. A detector score alone should never be treated as proof.

What should we do if my child is falsely accused of using AI?

Ask what evidence was used; if it is only a detector score, share the research on false positives and the policies of universities that turned detectors off. Version history, drafts and notes, plus having the student explain the work aloud, are the strongest evidence of authorship.

Does using AI make kids worse at thinking?

It can, depending on how it is used. Studies link heavy reliance to lower critical thinking and less brain engagement during writing, but most are small, correlational or preprints; the consistent pattern is that letting AI do the work hurts learning while using it to explain or question does not.

What does good AI use for homework look like?

Use it as a tutor or coach, not an answer machine: ask for hints, explanations, practice questions or feedback on your own draft, and do the first attempt yourself. In a Penn experiment, plain ChatGPT raised practice scores but lowered test scores, while a hint-only tutor avoided the drop.

Should my kid use the 'study mode' or 'learning mode' in chatbots?

These modes are designed to ask questions back instead of giving answers, which matches what research suggests helps learning. They are new, described mainly by the companies that sell them, and a student can still switch back to plain answers, so habits matter more than settings.

Does AI help or hurt creativity?

Both. In one experiment, stories written with AI ideas were rated more creative, especially for weaker writers, but they were more alike; another study found ChatGPT brainstorms less varied. Brainstorming alone first, then using AI, is a sensible approach.

What should my child's school AI policy say?

Good policies say which AI uses are allowed on each assignment, vet tools for privacy and benefit, teach AI literacy, and keep humans in charge of grades and discipline. Many schools still have no clear policy, and students without clear rules report avoiding or misusing AI.

How can teens learn to spot AI misinformation and fake images?

Looking closely at an image is not enough; studies show people mostly guess. The skill that works is lateral reading: leave the page, see what trusted sources say, and trace a claim to its origin, which a randomized classroom study showed teens can learn.

What do I need to know about deepfake nudes and sextortion?

AI 'nudify' apps and sextortion are real risks for teens, with financial sextortion mostly targeting boys. Tell your teen they will never be in trouble for coming to you, not to pay or comply, to keep evidence, and that Take It Down can help remove images taken under 18.

Will AI take the jobs my kid is training for? Does coding still matter?

Nobody knows for sure. Early data show fewer entry-level hires in some AI-exposed jobs, while employers expect demand for AI, tech and human skills like creative thinking to grow; computer science advocates argue understanding how AI works matters more, not less.

Where can my teen learn how AI actually works?

Several free, good options exist: Elements of AI and Crash Course AI for basics, Day of AI, Experience AI and Code.org for classroom-style lessons, and Harvard's CS50 AI for teens who already code.

The Big Index

AI at school and for homework

Older kids, essays and ChatGPT: thinking for yourself, detectors, media literacy and the future of work.

73 entries
65 free
10 must-reads ★

Questions parents ask

How many teens actually use AI for homework?
A lot, and the number is rising fast. Pew found teen ChatGPT use for schoolwork doubled from 13% (2023) to 26% (2024), and by fall 2025 just over half of teens had used a chatbot for schoolwork; RAND found 62% of students using AI for homework in late 2025.
- How Teens Use and View AI (Pew Research Center (McClain, Anderson, Sidoti & Bishop))
- About a quarter of U.S. teens have used ChatGPT for schoolwork – double the share in 2023 (Pew Research Center)
- RAND report on student AI use for homework (December 2025 survey) (RAND Corporation)
- The Dawn of the AI Era: Teens, Parents, and the Adoption of Generative AI at Home and School (Common Sense Media)
More in AI at school and for homework
Is using ChatGPT for homework cheating?
It depends on the assignment's rules and on whether AI does the thinking the task was meant to build. Teens themselves tend to see research help as fine and AI-written essays as not; the clearest fix is a teacher saying exactly how much AI help each assignment allows.
- About a quarter of U.S. teens have used ChatGPT for schoolwork – double the share in 2023 (Pew Research Center)
- The AI Assessment Scale (Perkins, Roe, Furze and MacVaugh)
- Common Sense Rules for the AI Era (Common Sense Education)
- Catch them Learning: A Pathway to Academic Integrity in the Age of AI (Tony Frontier, Cult of Pedagogy)
More in AI at school and for homework
Has AI made kids cheat more?
Not clearly. Stanford surveys found 60-70% of high schoolers admitted some cheating long before ChatGPT and the rate did not jump in 2023, though many teens now believe classmates use AI to cheat often. Cheating seems driven more by pressure and assignment design than by any single tool.
- What do AI chatbots really mean for students and cheating? (Stanford Graduate School of Education (Denise Pope, Victor Lee))
- How Teens Use and View AI (Pew Research Center (McClain, Anderson, Sidoti & Bishop))
- Catch them Learning: A Pathway to Academic Integrity in the Age of AI (Tony Frontier, Cult of Pedagogy)
More in AI at school and for homework
Can teachers tell if my kid used AI? Are AI detectors accurate?
Not reliably. Independent tests find detectors inaccurate and easy to fool, they flag non-native English writers far more often, OpenAI withdrew its own detector, and people's gut judgments are no better. A detector score alone should never be treated as proof.
- GPT detectors are biased against non-native English writers (Liang, Yuksekgonul, Mao, Wu and Zou, Stanford)
- Testing of Detection Tools for AI-Generated Text (Weber-Wulff et al.)
- OpenAI quietly shutters its AI text classifier due to low accuracy (Decrypt)
- Human heuristics for AI-generated language are flawed (Jakesch, Hancock and Naaman)
- Guidance on AI Detection and Why We're Disabling Turnitin's AI Detector (Vanderbilt University)
More in AI at school and for homework
What should we do if my child is falsely accused of using AI?
Ask what evidence was used; if it is only a detector score, share the research on false positives and the policies of universities that turned detectors off. Version history, drafts and notes, plus having the student explain the work aloud, are the strongest evidence of authorship.
- Guidance on AI Detection and Why We're Disabling Turnitin's AI Detector (Vanderbilt University)
- GPT detectors are biased against non-native English writers (Liang, Yuksekgonul, Mao, Wu and Zou, Stanford)
- Catch them Learning: A Pathway to Academic Integrity in the Age of AI (Tony Frontier, Cult of Pedagogy)
- GPTZero (GPTZero)
Does using AI make kids worse at thinking?
It can, depending on how it is used. Studies link heavy reliance to lower critical thinking and less brain engagement during writing, but most are small, correlational or preprints; the consistent pattern is that letting AI do the work hurts learning while using it to explain or question does not.
- Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task (Kosmyna et al., MIT Media Lab)
- AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking (Michael Gerlich, SBS Swiss Business School)
- The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers (Lee et al., Carnegie Mellon and Microsoft Research)
- Beware of Metacognitive Laziness: Effects of Generative Artificial Intelligence on Learning Motivation, Processes, and Performance (Fan, Tang ... Gašević)
- Against "Brain Damage" (Ethan Mollick, One Useful Thing)
What does good AI use for homework look like?
Use it as a tutor or coach, not an answer machine: ask for hints, explanations, practice questions or feedback on your own draft, and do the first attempt yourself. In a Penn experiment, plain ChatGPT raised practice scores but lowered test scores, while a hint-only tutor avoided the drop.
- Assigning AI: Seven Approaches for Students, with Prompts (Ethan Mollick and Lilach Mollick, Wharton)
- Kids who use ChatGPT as a study assistant do worse on tests (Jill Barshay, The Hechinger Report)
- AI Meets the Classroom: When Do Large Language Models Harm Learning? (Lehmann, Cornelius & Sting)
- Post-apocalyptic education (Ethan Mollick, One Useful Thing)
Should my kid use the 'study mode' or 'learning mode' in chatbots?
These modes are designed to ask questions back instead of giving answers, which matches what research suggests helps learning. They are new, described mainly by the companies that sell them, and a student can still switch back to plain answers, so habits matter more than settings.
- Introducing Claude for Education (Anthropic)
- Guided Learning in Gemini: From answers to understanding (Google)
- Kids who use ChatGPT as a study assistant do worse on tests (Jill Barshay, The Hechinger Report)
Does AI help or hurt creativity?
Both. In one experiment, stories written with AI ideas were rated more creative, especially for weaker writers, but they were more alike; another study found ChatGPT brainstorms less varied. Brainstorming alone first, then using AI, is a sensible approach.
- Generative artificial intelligence enhances creativity but reduces the diversity of novel content (Anil Doshi and Oliver Hauser)
- Homogenization Effects of Large Language Models on Human Creative Ideation (Anderson, Shah and Kreminski)
- Against "Brain Damage" (Ethan Mollick, One Useful Thing)
What should my child's school AI policy say?
Good policies say which AI uses are allowed on each assignment, vet tools for privacy and benefit, teach AI literacy, and keep humans in charge of grades and discipline. Many schools still have no clear policy, and students without clear rules report avoiding or misusing AI.
- AI Guidance for Schools Toolkit (TeachAI (Code.org, ETS, ISTE, Khan Academy, World Economic Forum))
- Common Sense Rules for the AI Era (Common Sense Education)
- State AI Guidance for K12 Schools (AI for Education)
- Walton Family Foundation-Gallup Voices of Gen Z: How American Youth View and Use Artificial Intelligence (Gallup and Walton Family Foundation)
How can teens learn to spot AI misinformation and fake images?
Looking closely at an image is not enough; studies show people mostly guess. The skill that works is lateral reading: leave the page, see what trusted sources say, and trace a claim to its origin, which a randomized classroom study showed teens can learn.
- Civic Online Reasoning (Digital Inquiry Group (Sam Wineburg and colleagues))
- Lateral Reading on the Open Internet (Wineburg et al., Digital Inquiry Group)
- SIFT (The Four Moves) (Mike Caulfield)
- A Representative Study on Human Detection of Artificially Generated Media Across Countries (Frank et al., CISPA and Ruhr University Bochum)
- Checkology (News Literacy Project)
What do I need to know about deepfake nudes and sextortion?
AI 'nudify' apps and sextortion are real risks for teens, with financial sextortion mostly targeting boys. Tell your teen they will never be in trouble for coming to you, not to pay or comply, to keep evidence, and that Take It Down can help remove images taken under 18.
- Deepfake Nudes & Young People: Navigating a New Frontier in Technology-facilitated Nonconsensual Sexual Abuse and Exploitation (Thorn)
- Grooming and sextortion (Thorn research) (Thorn)
- Take It Down (National Center for Missing & Exploited Children)
- Sextortion (NCMEC) (National Center for Missing & Exploited Children)
Will AI take the jobs my kid is training for? Does coding still matter?
Nobody knows for sure. Early data show fewer entry-level hires in some AI-exposed jobs, while employers expect demand for AI, tech and human skills like creative thinking to grow; computer science advocates argue understanding how AI works matters more, not less.
- Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence (Brynjolfsson, Chandar and Chen, Stanford Digital Economy Lab)
- Future of Jobs Report 2025 (World Economic Forum)
- CSforALL (CSforALL)
- AI as Normal Technology (Sayash Kapoor and Arvind Narayanan)
Where can my teen learn how AI actually works?
Several free, good options exist: Elements of AI and Crash Course AI for basics, Day of AI, Experience AI and Code.org for classroom-style lessons, and Harvard's CS50 AI for teens who already code.
- Elements of AI (University of Helsinki & MinnaLearn)
- Crash Course Artificial Intelligence (Crash Course (with PBS Digital Studios))
- Day of AI (MIT RAISE)
- Experience AI (Raspberry Pi Foundation & Google DeepMind)
- CS50's Introduction to Artificial Intelligence with Python (Harvard CS50 (Brian Yu, David Malan))

Must-reads

ArticleFree
Kids who use ChatGPT as a study assistant do worse on tests
Jill Barshay, The Hechinger Report, 2024
Explains a Penn experiment with nearly 1,000 high schoolers in Turkey. Plain ChatGPT boosted practice scores but students then did worse on the test; a hint-only AI tutor avoided that drop. One study, then a working paper.
hechingerreport.org
StudyFree
Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task
Kosmyna et al., MIT Media Lab, 2025
Much-discussed preprint: 54 students wrote essays with ChatGPT, search, or nothing while wearing EEG caps. The ChatGPT group showed the least brain engagement and could barely quote their own essays. Small, college-age and debated.
arxiv.org
CourseFree
AI Literacy Lessons for Grades K–12
Common Sense Education
Free short lessons on AI bias, chatbots and friendship, deepfakes and consent, creativity and ownership, and checking AI answers with the SIFT method. Easy to use at the kitchen table as conversation starters too.
commonsense.org
ReportFree
RAND report on student AI use for homework (December 2025 survey)
RAND Corporation, 2026
Survey of about 1,200 students from middle school through college. 62% used AI for homework, and two-thirds agreed it hurts critical thinking, yet kept using it. Recommends schools separate AI that does the thinking from AI that stretches it.
rand.org
GuideFree
Assigning AI: Seven Approaches for Students, with Prompts
Ethan Mollick and Lilach Mollick, Wharton, 2023
Widely used working paper describing seven roles AI can play for a student, such as tutor, coach, teammate or simulator, each with a ready-to-copy prompt and its risks. Core idea: the student stays in charge of the thinking.
papers.ssrn.com
StudyFree
GPT detectors are biased against non-native English writers
Liang, Yuksekgonul, Mao, Wu and Zou, Stanford, 2023
Stanford study, published in Patterns, showing popular AI detectors flagged most essays by non-native English writers as AI-written while judging U.S. eighth graders' essays correctly. Simple rewording fooled the detectors. Authors advise against using them to grade students.
arxiv.org
StudyFree
Generative artificial intelligence enhances creativity but reduces the diversity of novel content
Anil Doshi and Oliver Hauser, 2024
Experiment published in Science Advances. Short stories written with AI ideas were rated more creative, especially for weaker writers, but they were more alike. A useful frame: AI can lift individuals while flattening the group.
arxiv.org
GuideFree
AI Guidance for Schools Toolkit
TeachAI (Code.org, ETS, ISTE, Khan Academy, World Economic Forum), 2023
The most-used free template for school AI policies, updated in 2025. Seven principles, sample policy language, family engagement tips and a list of risks like privacy, bias and overreliance. Shows parents what a good school policy looks like.
teachai.org
CourseFree
Civic Online Reasoning
Digital Inquiry Group (Sam Wineburg and colleagues)
Free lessons, assessments and videos built on how professional fact checkers work: who is behind this, what is the evidence, what do other sources say. Start with the lateral reading and click restraint lessons.
cor.inquirygroup.org
ToolFree
Take It Down
National Center for Missing & Exploited Children, 2023
Free, anonymous service that helps remove or block nude or sexual images of someone taken before age 18, including on participating platforms, without sharing the image itself. Bookmark it before you ever need it.
takeitdown.ncmec.org

All 73

Using AI well for homework

★Kids who use ChatGPT as a study assistant do worse on testsJill Barshay, The Hechinger Report · 2024 · Article · Free · TeensExplains a Penn experiment with nearly 1,000 high schoolers in Turkey. Plain ChatGPT boosted practice scores but students then did worse on the test; a hint-only AI tutor avoided that drop. One study, then a working paper.
★Assigning AI: Seven Approaches for Students, with PromptsEthan Mollick and Lilach Mollick, Wharton · 2023 · Guide · Free · TeensWidely used working paper describing seven roles AI can play for a student, such as tutor, coach, teammate or simulator, each with a ready-to-copy prompt and its risks. Core idea: the student stays in charge of the thinking.
How to Use ChatGPT as an Example MachineChris Mah and Sarah Levine, Cult of Pedagogy · 2023 · Podcast · Free · Ages 8–12Podcast episode and post from two Stanford education researchers. Use AI to quickly make contrasting examples and non-examples for students to compare, a proven learning technique, while checking the output for errors.
Artificial Intelligence Teaching GuideStanford Teaching Commons · 2023 · Guide · FreeSix free modules for instructors: AI literacy, how chatbots work, classroom uses and risks, writing an AI policy for a syllabus, and redesigning assignments. Written for college but easy to adapt for high school.
AI Pedagogy ProjectmetaLAB at Harvard · 2023 · Guide · Free · TeensFree collection of classroom assignments that use AI and ask students to reflect on its strengths and limits, plus a beginner guide to chatbots and sample policies. Made at Harvard's Berkman Klein Center.

Thinking, offloading and creativity

★Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing TaskKosmyna et al., MIT Media Lab · 2025 · Study · Free · TeensMuch-discussed preprint: 54 students wrote essays with ChatGPT, search, or nothing while wearing EEG caps. The ChatGPT group showed the least brain engagement and could barely quote their own essays. Small, college-age and debated.
AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical ThinkingMichael Gerlich, SBS Swiss Business School · 2025 · Study · FreeSurvey and interviews with 666 people in the journal Societies. Heavier AI use went with lower critical-thinking scores, especially among younger people, linked to handing off mental work. Correlational only, so it cannot show AI causes the drop.
The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge WorkersLee et al., Carnegie Mellon and Microsoft Research · 2025 · Study · FreeCHI 2025 survey of 319 workers. The more people trusted AI, the less they reported thinking critically; the more they trusted themselves, the more they checked. Adults, self-reported, but a useful lesson for teens on staying the checker.
Does using AI dumb you down?WBUR On Point (Meghna Chakrabarti) · 2025 · Podcast · FreePublic radio hour with the MIT study lead and a Johns Hopkins neurologist on the essay-writing EEG study and what it may mean for students. An easy listen that explains cognitive debt and where the evidence is thin.
★Generative artificial intelligence enhances creativity but reduces the diversity of novel contentAnil Doshi and Oliver Hauser · 2024 · Study · FreeExperiment published in Science Advances. Short stories written with AI ideas were rated more creative, especially for weaker writers, but they were more alike. A useful frame: AI can lift individuals while flattening the group.
Homogenization Effects of Large Language Models on Human Creative IdeationAnderson, Shah and Kreminski · 2024 · Study · FreeSmall study of 36 people brainstorming. ChatGPT users produced more and more detailed ideas, but ideas were less varied across people and users felt less ownership of them. Supports brainstorming on your own first.

School AI policy and AI literacy

★AI Literacy Lessons for Grades K–12Common Sense Education · Course · Free · TeensFree short lessons on AI bias, chatbots and friendship, deepfakes and consent, creativity and ownership, and checking AI answers with the SIFT method. Easy to use at the kitchen table as conversation starters too.
★AI Guidance for Schools ToolkitTeachAI (Code.org, ETS, ISTE, Khan Academy, World Economic Forum) · 2023 · Guide · FreeThe most-used free template for school AI policies, updated in 2025. Seven principles, sample policy language, family engagement tips and a list of risks like privacy, bias and overreliance. Shows parents what a good school policy looks like.
TeachAI Policy TrackerTeachAI · Data · FreeRunning collection of official AI-in-education guidance from U.S. states and several countries, including Australia, Canada, Japan and the UK. Useful to see what your state or country has said.
State AI Guidance for K12 SchoolsAI for Education · 2026 · Data · FreeRegularly updated list summarizing each state's official K-12 AI guidance with links to the documents. As of August 2026 it lists 35 states plus Puerto Rico with formal guidance.
AI for EducationAI for Education (Amanda Bickerstaff) · Organization · Free + paidTraining group that helps schools adopt generative AI. Offers a free prompt library, a free two-hour course for educators, free student lessons and guides on detectors and policy; its school training is paid.
ISTE+ASCD Artificial Intelligence in EducationISTE+ASCD · Organization · Some freeHub from the largest ed-tech teacher association: free lesson ideas, a profile of an AI-ready graduate, leader guides, courses and books. Some items are free, courses and books cost money.
Commonsense Guardrails for Using Advanced Technology in SchoolsAmerican Federation of Teachers · 2024 · Guide · FreeThe teachers union's guide to AI in schools, focused on keeping teachers in charge, protecting student data and judging tools before buying. Shows how many educators want AI adopted: slowly and with guardrails.
AI competency framework for studentsUNESCO · 2024 · Guide · FreeTwelve AI skills students should build, from a human-centered mindset and ethics to how AI systems work and how to design them, in three levels. Helps parents see what real AI literacy covers beyond typing prompts.
Common Sense Rules for the AI EraCommon Sense Education · 2026 · Guide · FreeTen plain rules for schools: teach AI literacy at every grade, vet tools for safety and proven benefit, say which AI uses are allowed on each assignment, design work AI cannot shortcut, and tell families what tools collect.

Surveys: who uses AI for school

About a quarter of U.S. teens have used ChatGPT for schoolwork – double the share in 2023Pew Research Center · 2025 · Report · Free · TeensShort Pew read from a fall 2024 survey: 26% of teens used ChatGPT for schoolwork, up from 13%. Most teens think using it to research a topic is fine, but few accept using it to write an essay.
About 1 in 5 U.S. teens who've heard of ChatGPT have used it for schoolworkPew Research Center · 2023 · Report · Free · TeensThe first big baseline, from fall 2023. Older teens used ChatGPT for school more often, and teens drew a clear line: research help seemed acceptable, essay writing mostly did not. Useful for seeing how fast things changed.
A quarter of U.S. teachers say AI tools do more harm than good in K-12 educationPew Research Center · 2024 · Report · FreeSurvey of 2,531 public school teachers. High school teachers were the most negative about AI tools, very few at any level said they do more good than harm, and many elementary teachers were simply unsure.
★RAND report on student AI use for homework (December 2025 survey)RAND Corporation · 2026 · Report · Free · TeensSurvey of about 1,200 students from middle school through college. 62% used AI for homework, and two-thirds agreed it hurts critical thinking, yet kept using it. Recommends schools separate AI that does the thinking from AI that stretches it.
See Which Types of Teachers Are the Early Adopters of AIEducation Week · 2024 · Article · Free + paidNews write-up of RAND's 2024 teacher panel data on which teachers were first to use AI for planning and instruction. Good background on how uneven classroom use was early on. May hit a paywall after a few articles.
The AI Dividend: New Survey Shows AI Is Helping Teachers Reclaim Valuable TimeGallup and Walton Family Foundation · 2025 · Report · FreeSurvey of 2,232 public school teachers. Weekly AI users report saving almost six hours a week, mostly on lesson prep and worksheets, but only 19% say their school has an AI policy. The foundation funds AI-in-education work.
Walton Family Foundation-Gallup Voices of Gen Z: How American Youth View and Use Artificial IntelligenceGallup and Walton Family Foundation · 2025 · Report · Free · TeensSurvey of 3,465 people ages 13-28. Most have used AI, anxiety beats excitement, and nearly half of students avoided AI for schoolwork because they did not know if it was allowed. Clear school rules went with feeling better prepared.

AI detectors and integrity

★GPT detectors are biased against non-native English writersLiang, Yuksekgonul, Mao, Wu and Zou, Stanford · 2023 · Study · Free · TeensStanford study, published in Patterns, showing popular AI detectors flagged most essays by non-native English writers as AI-written while judging U.S. eighth graders' essays correctly. Simple rewording fooled the detectors. Authors advise against using them to grade students.
Testing of Detection Tools for AI-Generated TextWeber-Wulff et al. · 2023 · Study · FreeEuropean academic-integrity researchers tested 14 detectors, including Turnitin. None was accurate or reliable, they leaned toward calling text human, and light editing made them worse. Published in the International Journal for Educational Integrity.
Can AI-Generated Text be Reliably Detected?Sadasivan et al., University of Maryland · 2023 · Study · FreeTechnical paper showing that paraphrasing AI text defeats every major kind of detector, including watermarks, and that human writing can be made to look AI-made. Argues detection gets harder as models improve.
A Watermark for Large Language ModelsKirchenbauer et al., University of Maryland · 2023 · Study · FreeThe best-known proposal for hiding a statistical signal in AI-written text so software can later spot it. Useful background on why watermarks are discussed as an alternative to guess-based detectors, and why they only work if the AI company adds them.
Human heuristics for AI-generated language are flawedJakesch, Hancock and Naaman · 2023 · Study · FreeSix experiments with 4,600 people, published in PNAS. People could not tell AI writing from human writing and relied on wrong cues like contractions or family mentions. A reminder that teachers' gut feelings are not a detector either.
Guidance on AI Detection and Why We're Disabling Turnitin's AI DetectorVanderbilt University · 2023 · Guide · FreeA university explains turning off Turnitin's AI detector: even a 1% false positive rate could wrongly flag hundreds of papers a year, plus bias against non-native writers and no transparency. Handy to share if a child is falsely accused.
OpenAI quietly shutters its AI text classifier due to low accuracyDecrypt · 2023 · Article · FreeNews report on ChatGPT's maker withdrawing its own AI-writing detector in July 2023. In OpenAI's tests it caught only about a quarter of AI text and wrongly flagged some human writing. Historically important; the tool is gone.
GPTZeroGPTZero · 2023 · Tool · Free + paid · TeensOne of the most-used AI writing detectors, free for short checks with paid plans. Its accuracy claims come from the company itself; independent studies find detectors unreliable, and the site says results should not be used alone to punish anyone.
What do AI chatbots really mean for students and cheating?Stanford Graduate School of Education (Denise Pope, Victor Lee) · 2023 · Article · Free · TeensLong-running surveys show 60-70% of high schoolers admitted some cheating before ChatGPT existed, and the 2023 rate did not jump. Useful perspective: cheating has deeper causes than any single tool.
Catch them Learning: A Pathway to Academic Integrity in the Age of AITony Frontier, Cult of Pedagogy · 2025 · Podcast · Free · TeensPodcast episode and essay arguing that catching cheaters fails when detection tools lag. Instead, build integrity by having students show and explain their process, with checkpoints and clear AI rules per assignment.
The AI Assessment ScalePerkins, Roe, Furze and MacVaugh · 2023 · Tool · Free · TeensA five-level scale teachers can attach to any assignment, from 'no AI' to 'AI exploration'. Popular because it replaces guessing with clear expectations, so students know exactly how much AI help is allowed.

Courses for teens

AI, Effort, and Your BrainDay of AI, MIT RAISE · Course · Free · TeensAbout an hour-long lesson for ages 11-18 on cognitive effort and 'cognitive debt': when AI helps you learn and when it does the thinking for you. Ends with class norms for AI use. Full materials need a login.
CS50's Introduction to Artificial Intelligence with PythonHarvard CS50 (Brian Yu, David Malan) · 2020 · Course · Free · TeensFree Harvard course on search, machine learning and language models, with real coding projects. Demanding: it expects about a year of Python. Best for older teens who already code and want to build AI, not just use it.

Media literacy and misinformation

Truth, Tricks, and AI: Learning to Verify InformationDay of AI, MIT RAISE · Course · Free · TeensHour-long lesson for ages 11 and up on AI hallucinations, deepfakes and biased answers. Students practice lateral reading to check claims. Key point: AI can sound sure of itself while being wrong.
News Literacy ProjectNews Literacy Project · Organization · Free · TeensNonpartisan nonprofit that teaches students to tell credible information from misinformation. Runs Checkology, the Sift newsletter for teachers, National News Literacy Week and a free family guide.
CheckologyNews Literacy Project · Course · Free · TeensFree online classroom with more than 250 interactive lessons on misinformation, bias, sourcing and algorithms, including generative AI. Designed for grades 5-12; no paid tier.
RumorGuardNews Literacy Project · Tool · Free · TeensNews Literacy Project's site that breaks down viral rumors and fake images and shows how to check them yourself. Handy for practicing with a teen on real examples from the news feed.
★Civic Online ReasoningDigital Inquiry Group (Sam Wineburg and colleagues) · Course · Free · TeensFree lessons, assessments and videos built on how professional fact checkers work: who is behind this, what is the evidence, what do other sources say. Start with the lateral reading and click restraint lessons.
Lateral Reading on the Open InternetWineburg et al., Digital Inquiry Group · 2022 · Study · Free · TeensRandomized classroom study in an urban district, published in the Journal of Educational Psychology. Teens taught to leave a website and check what others say about it got much better at judging online credibility.
SIFT (The Four Moves)Mike Caulfield · 2019 · Blog · Free · TeensThe original post on SIFT: Stop, Investigate the source, Find better coverage, Trace claims to the original. A simple routine teens can apply to posts, images and chatbot answers alike.
MediaWisePoynter Institute · 2018 · Organization · Free · TeensPoynter's media and AI literacy project. For teens it offers AI Unlocked, a five-part video series on using generative AI safely, a free Digital Media Literacy 101 course and a Teen Creator Network of fact checkers ages 13-20.
Calling Bullshit: Data Reasoning in a Digital WorldCarl Bergstrom and Jevin West, University of Washington · 2017 · Course · Free · TeensUniversity course materials, videos and case studies on spotting misleading statistics, charts and models, plus a companion book. The title includes a mild swear word; the content suits older teens.

Future of work and careers

Hour of AICSforALL and Code.org · 2025 · Course · FreeSuccessor to the Hour of Code: free one-hour hands-on AI activities for every age from partners like Google, Microsoft, LEGO Education and Scratch, run around Computer Science Education Week each December and all year.
CSforALLCSforALL · Organization · FreeNonprofit pushing for computer science and AI education for every student; co-runs Hour of AI. Its 'CS Is Everything' campaign argues coding still matters for art, music, sports and activism, not only tech jobs.
Google AI EssentialsGoogle, via Coursera · 2024 · Course · Paid · TeensShort self-paced course on using AI tools for everyday work tasks, with a certificate. Aimed at adults and job seekers rather than kids, and it needs a paid Coursera subscription after a free trial.
AI as Normal TechnologySayash Kapoor and Arvind Narayanan · Newsletter · FreeFree newsletter from the AI Snake Oil authors arguing AI is a big but ordinary technology that will change jobs gradually, not overnight. A level-headed counterweight to hype about kids' future careers.
Future of Jobs Report 2025World Economic Forum · 2025 · Report · Free · TeensBig employer survey on jobs through 2030. AI and big data, cybersecurity and tech literacy are the fastest-growing skills, alongside creative thinking, resilience and curiosity. Large PDF; employer forecasts, not certainties.
Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial IntelligenceBrynjolfsson, Chandar and Chen, Stanford Digital Economy Lab · 2025 · Study · Free · TeensPayroll data show jobs for 22-25-year-olds in AI-exposed fields like software and customer service running well below expectations, mostly from fewer hires. Jobs where AI helps rather than replaces held up. Working paper, descriptive not causal.

Teacher voices, books and blogs

Teaching with AI: A Practical Guide to a New Era of Human Learning (second edition)José Antonio Bowen and C. Edward Watson · Book · PaidPopular practical book for educators on designing assignments, policies and feedback when students have AI. Written mainly for college faculty. The authors run a free companion site with slides and starter guides.
We Teach With AIJosé Antonio Bowen and C. Edward Watson · Guide · FreeFree companion site to the book Teaching with AI, with a start-here guide, a list of current AI tools, design principles for learning with AI and the authors' presentation slides.
The AI ClassroomDan Fitzpatrick and Amanda Fox · 2023 · Book · PaidSite for the K-12 teacher book The AI Classroom by Dan Fitzpatrick and colleagues, with a newsletter and paid training. A hands-on, enthusiastic take on using AI for lesson planning and student projects.
AI Snake OilArvind Narayanan and Sayash Kapoor, Princeton · 2024 · Book · Paid · TeensTwo Princeton computer scientists sort AI that works from AI that does not, including in education and hiring. Builds the healthy skepticism teens need, without doom. Great for a debate or reading club.
The Important WorkJane Rosenzweig · Newsletter · Free · TeensNewsletter on teaching writing in the age of AI, from the director of a Harvard writing center. Thoughtful on why writing is thinking and what is lost when students hand it off.
RhetoricaMarc Watkins · Newsletter · Free · TeensNewsletter from a university writing instructor who trains teachers on AI. Practical and skeptical in equal measure, covering AI writing tools, detection, and what students say they actually do.
Leon Furze (blog)Leon Furze · Blog · FreeAustralian educator and co-creator of the AI Assessment Scale writing frequently about school AI policy, deepfakes, privacy, AI agents and what teachers actually do with AI. Practical, critical and current.
The Biblioracle RecommendsJohn Warner · Newsletter · Free · TeensWeekly essays from a longtime writing teacher and author of More Than Words, a book on why writing must stay human. Often returns to AI, reading and what school writing assignments are for.
I Quit Teaching Because of ChatGPTVictoria Livingstone, TIME · 2024 · Article · Free · TeensOpinion essay by a longtime writing instructor who left teaching after students leaned on AI to write and paraphrase. Argues writing is how we think, and she was grading AI more than students. A strong statement of the worry.
An AI Wish List From Teachers: What They Actually Want It to DoEllen Ullman, EdSurge · 2025 · Article · FreeTeachers describe wanting AI for routine work like rubrics, translations and leveled readings, while keeping grading decisions and relationships human. Shows how classroom AI looks from the teacher's side.

Deepfakes, nudify apps and sextortion

Which Face Is Real?Jevin West and Carl Bergstrom, University of Washington · 2019 · Game · Free · Ages 8–12Quick game: pick the real photo from a pair where one face was made by AI. Fun to play together and a fast way to show teens how convincing fake faces have become.
A Representative Study on Human Detection of Artificially Generated Media Across CountriesFrank et al., CISPA and Ruhr University Bochum · 2023 · Study · FreeSurvey of 3,002 adults in the U.S., Germany and China (IEEE S&P 2024). People were basically guessing when judging AI images, audio and text, and tended to call fakes real. Spotting by eye is not a dependable defense.
Content CredentialsCoalition for Content Provenance and Authenticity (C2PA) · Tool · FreeAn industry standard that attaches a 'nutrition label' to images and video showing how they were made and edited, backed by Adobe, Microsoft, Google, OpenAI and others. Helpful to know, but most content online still carries no label.
SynthIDGoogle DeepMind · 2023 · Tool · FreeGoogle's invisible watermark for AI images, video, audio and text, with a checker you can use in Gemini. It only finds content from Google and partner tools, so a 'no watermark' result does not prove something is real.
Deepfake Nudes & Young People: Navigating a New Frontier in Technology-facilitated Nonconsensual Sexual Abuse and ExploitationThorn · 2025 · Report · Free · TeensSurvey research from a child-safety nonprofit on AI 'nudify' images among young people. Most teens see deepfake nudes as abuse, the problem is already real in schools, and the apps are easy to reach. Grounds a calm family talk.
Grooming and sextortion (Thorn research)Thorn · Report · Free · TeensResearch summary on sextortion. Financial sextortion mostly targets teen boys and can escalate within hours; paying rarely ends threats; some threats now use fake or AI-made images. Useful facts before talking with a teen.
★Take It DownNational Center for Missing & Exploited Children · 2023 · Tool · Free · TeensFree, anonymous service that helps remove or block nude or sexual images of someone taken before age 18, including on participating platforms, without sharing the image itself. Bookmark it before you ever need it.
Sextortion (NCMEC)National Center for Missing & Exploited Children · Guide · Free · TeensPlain guide to sextortion: red flags, why not to pay, how to report and block without deleting evidence, and where to get help, including threats to make fake explicit images. Links videos and a parent tip sheet.