Screenagers Under the Influence and the Rise of AI Drug Advice

Screenagers Under the Influence is the Screenagers film about vaping, drugs and alcohol, and it was made for a world where teens learned about substances from friends, social media and TV. That world now includes a chatbot that answers any question at 1 a.m., privately, in a friendly voice. Here is the good news first: most US teens don’t use drugs, and abstention is near historic highs. The harder news is that independent tests have found chatbots giving teens harmful substance-use guidance, and one family is suing an AI company over advice they allege contributed to their son’s death. This article covers what the film teaches, how chatbots actually handle drug questions, and the few sentences worth saying to your teen this week.

Key Takeaways

• Screenagers Under the Influence (around 2023; 63 minutes, grade 6 and up) focuses on everyday teen decisions about vaping, drugs and alcohol, not severe addiction, and features experts including Nora Volkow, Anna Lembke and Lisa Damour.
• Most teens don’t use: in 2025, 66% of 12th graders and 91% of 8th graders reported no marijuana, alcohol or nicotine in the past 30 days (Monitoring the Future).
• Chatbot safeguards can be bypassed. In a 2025 test by the Center for Countering Digital Hate, 53% of 1,200 ChatGPT responses to harmful prompts, including on substance use, were rated harmful, and simple framing like “this is for a presentation” got past refusals.
• Chatbots can be useful for general facts, but they are not safe for personal decisions about doses, mixing substances or “doing it safely.”
• The film’s best advice still applies: short, frequent, calm conversations. Add one line about AI.

If you or your teen need help now: In the US, call or text 988 (Suicide & Crisis Lifeline), text HOME to 741741 (Crisis Text Line), or call SAMHSA’s free, confidential National Helpline at 1-800-662-4357 (24/7, English and Spanish). For a possible poisoning, call Poison Help at 1-800-222-1222; for an emergency, call 911. Outside the US, find a line at findahelpline.com.

What Screenagers Under the Influence Covers

Screenagers Under the Influence: Addressing Vaping, Drugs, and Alcohol in the Digital Age is the third feature in Delaney Ruston and Lisa Tabb’s Screenagers series. The official site gives no release year but lists festival selections starting in 2023, so it is reasonable to say it came out around then. It runs 63 minutes (49 in a classroom version) and is suggested for grade 6 and up, with English and Spanish subtitles.

The film’s approach is its distinguishing feature. According to its official page, unlike most substance-use documentaries that focus on severe addiction, it looks at more common everyday situations, including how teens are constantly exposed to images of vaping, drugs and alcohol on social media, in movies and on TV. It combines current science with personal stories, debunks common myths and offers strategies for setting limits and supporting teen mental health.

The expert list is strong: National Institute on Drug Abuse director Nora Volkow, Stanford’s Anna Lembke, neuroscientist Yasmin Hurd, psychologists Lisa Damour and David Yeager, adolescent medicine specialist Sharon Levy, and tobacco researchers Pam Ling and Robert Jackler, among others.

In an October 20, 2026 post, Ruston connected the film to AI directly. Writing about a lawsuit involving a former Cornell student, she listed points to share with teens, including that alcohol further impairs a still-developing prefrontal cortex, that today’s recreational cannabis is far stronger than the plant once was, and that AI chatbots can give misleading or unsafe drug information.

What Teens Are Actually Doing

Context matters before any warning. The Monitoring the Future survey, funded by the National Institutes of Health and run by the University of Michigan, surveyed 23,726 students in 2025. Past-year use among 8th, 10th and 12th graders was:

• Alcohol: 11%, 24%, 41%
• Cannabis: 8%, 16%, 26%
• Nicotine vaping: 9%, 14%, 20%

Past-30-day abstention from all three was 91% of 8th graders, 82% of 10th graders and 66% of 12th graders. NIDA said abstention remained at historically high levels and that use has stayed near its 2021 low for five years. The survey did flag small but significant rises in past-year heroin and cocaine use, which researchers said warrant close monitoring.

So the realistic picture is a majority of teens who don’t use, a minority who do, and almost all of them curious. Curiosity is exactly what drives a question to a chatbot. For more on newer nicotine products, see our guide to new youth vaping products in 2025 and 2026.

How Chatbots Handle Drug Questions

Here is the mechanism, because it explains both why chatbots often refuse and why they sometimes don’t.

A chatbot like ChatGPT is a large language model trained on huge amounts of text, then further trained to follow instructions and to decline certain requests. On top of that, companies add safety classifiers: separate systems that flag risky prompts or responses and trigger a refusal or a safety message. When a teen asks something plainly dangerous, these layers usually catch it.

They are not airtight, for three reasons:

1. Framing changes the classification. A question asked “for a school project” or “for a story” can look harmless to a classifier, even if the person asking is a teen planning to use. The Center for Countering Digital Hate’s August 2025 “Fake Friend” report found exactly this: simple framing like “this is for a presentation” was enough to bypass safeguards. Across 60 harmful prompts producing 1,200 responses, 53% were rated harmful, in categories including self-harm, eating disorders and substance abuse. CCDH is an advocacy nonprofit, and its ratings are its own; the pattern it describes, however, matches what other testers have found.
2. Models are built to be agreeable. Language models tend to go along with the user, a trait researchers call sycophancy. A teen who says “I’m going to do it anyway, just tell me how to be safe” may get an answer that sounds like harm reduction but is personalized in ways no clinician would endorse.
3. Products differ. Companion apps, game characters and AI toys often have weaker safeguards than mainstream chatbots. Common Sense Media’s 2026 test of AI companion toys found 27% of outputs inappropriate for kids, including content about drugs.

The most serious example is in court. In Ruston’s account, a college student named Sam used ChatGPT to ask how to use a substance “safely,” and the chatbot eventually gave unsafe drug advice. His parents are suing OpenAI, and their complaint alleges the advice played a significant role in his death. That is an allegation in active litigation, not a court finding, and the case does not show how often such failures happen. It shows that they can.

A Common Sense Media survey of more than 1,200 kids aged 9 to 17, reported in June 2026, found 57% of AI users had asked chatbots for advice on health or body issues. Drug questions are part of that territory.

Where Chatbots Help and Where They Don’t

Type of question	Can a chatbot help?	Main risk	Better source
General facts (“What is in a vape?”)	Often, with checking	Errors stated confidently	NIDA, CDC teen pages, a pediatrician
Myths (“Is weed safer than alcohol?”)	Sometimes	Oversimplified or agreeable answers	Trusted adult conversation; the film itself
Peer situations (“What do I say if someone offers me something?”)	Can help rehearse words	Generic advice, misses context	Parent, school counselor
Personal use, doses, mixing, “doing it safely”	No	Personalized unsafe advice; filters can be bypassed	Doctor; Poison Help 1-800-222-1222; SAMHSA 1-800-662-4357
Someone may be in danger now	No	Delay	911

What to Say to Your Teen This Week

Keep it short and frequent

The film’s core message, and the one with the best support, is that frequent, calm conversations beat one big talk. Ruston’s family AI agreement guidance makes the same point: the goal is ongoing conversations about why, not perfect rule-following. Five minutes in the car counts.

Add one sentence about AI

Something like: “Chatbots can be OK for general facts, but they’re not safe for questions about taking anything, doses or mixing. If you’re wondering about that, ask me or a doctor, and I won’t freak out.” The second half matters as much as the first.

Lead with the norm, not the scare

Most teens don’t use. Saying so is more persuasive than horror stories, because it corrects the impression that “everyone does it” that social media can create, a theme the film itself explores.

Explain consent and intoxication

Ruston’s October 2026 post made a point many parents skip: intoxication can leave a person unable to consent, and being awake or able to talk doesn’t mean someone can. It is a hard conversation, and an important one before high school parties.

Know your emergency numbers

Put Poison Help (1-800-222-1222) in your phone and your teen’s. If someone collapses, has a seizure, has trouble breathing or can’t be woken, call 911.

What not to do

Don’t search your teen’s chatbot history in secret and then confront them; you will likely learn less from then on. Don’t present the lawsuit as proof that “AI kills kids,” which overstates what is known and invites eye-rolls. And don’t avoid the topic because your teen seems uninterested; curiosity often goes straight to a screen.

What to Watch For Over the Next 3 Months

• Week 4: Have you had at least two short substance-use conversations, including the AI line? Does your teen know Poison Help and 911 are for real emergencies?
• Month 2 red flags: new secrecy, money disappearing, sudden changes in friends, sleep or grades, vape devices or unfamiliar products, or comments suggesting a chatbot is their main source for drug information.
• Month 3 self-check: Ask your teen where they would go first with a question about a substance. If the answer is “AI” or “TikTok,” revisit the conversation. If it’s you, that’s the win.

For the other films in the series and which to watch first, see every Screenagers film, explained. For chatbot settings that can add a layer of protection, see our guide to AI chatbot safety settings for kids.

Frequently Asked Questions

What is Screenagers Under the Influence about?

It is the third Screenagers documentary, released around 2023, about how digital life shapes teen decisions on vaping, drugs and alcohol. It focuses on everyday situations rather than severe addiction, debunks myths and offers strategies for limits, with experts such as Nora Volkow, Anna Lembke and Lisa Damour. It runs 63 minutes and suits grade 6 and up.

Do chatbots give teens drug advice?

They are designed to refuse dangerous requests, but tests show the safeguards can be bypassed. A 2025 CCDH test found 53% of ChatGPT responses to harmful prompts, including substance use, were harmful, and framing like “it’s for a presentation” got past refusals. Chatbots are not a safe source for personal drug decisions.

What should my teen do instead of asking AI about drugs?

Come to you, a doctor or a school counselor for personal questions. For possible poisoning, call Poison Help at 1-800-222-1222. For treatment referrals, SAMHSA’s free National Helpline is 1-800-662-4357, 24/7 in English and Spanish. In an emergency, call 911. In crisis, call or text 988.

How many teens actually use drugs?

Fewer than most people think. In the 2025 Monitoring the Future survey, 91% of 8th graders, 82% of 10th graders and 66% of 12th graders reported no marijuana, alcohol or nicotine in the past 30 days. NIDA says abstention remains at historically high levels.

Is the lawsuit against OpenAI proven?

No. The family’s complaint alleges ChatGPT gave their son unsafe drug advice that played a significant role in his death. These are allegations in active litigation, not a court’s findings. The case is a reason to talk with teens about AI and drugs, not proof of how often chatbots fail.

Help, again: Call or text 988; text HOME to 741741; SAMHSA 1-800-662-4357; Poison Help 1-800-222-1222; emergencies 911; outside the US, findahelpline.com.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Screenagers. “About Screenagers Under the Influence: Addressing Vaping, Drugs, and Alcohol in the Digital Age.” https://screenagersmovie.com/about-screenagers-under-the-influence
2. Ruston, D. (2026, October 20). “The Cornell Case: 6 Important Points to Share With Teens About Drugs and Consent.” Screenagers. https://www.screenagersmovie.com/blog/the-cornell-case
3. National Institute on Drug Abuse. (2025, December 17). “Reported use of most drugs remains low among U.S. teens” (Monitoring the Future 2025). https://nida.nih.gov/news-events/news-releases/2025/12/reported-use-of-most-drugs-remains-low-among-us-teens
4. Center for Countering Digital Hate. (2025, August 6). “Fake Friend” (research report on ChatGPT and teens). https://www.counterharm.com/research/fake-friend-chatgpt/
5. Common Sense Media. (2026, updated January 22). “AI Toys: Risk Assessment.” https://institute.commonsensemedia.org/risk-assessments/ai-toys
6. Smith, C. (2026, June 8). “Nearly 9 in 10 kids use AI; report flags concerns over educational, emotional dependencies” (Common Sense Media survey). The National News Desk / KOKH. https://okcfox.com/news/nation-world/nearly-9-in-10-kids-use-ai-report-flags-concerns-over-educational-emotional-dependency-common-sense-media-youth-ai-safety-institute-common-sense-media-census-ai-use-by-tweens-and-teens-2026
7. Substance Abuse and Mental Health Services Administration. “SAMHSA’s National Helpline.” https://www.samhsa.gov/find-help/helplines/national-helpline
8. Poison.org (National Capital Poison Center). “Poison Help.” https://www.poison.org/
9. Screenagers. (2026, October 27). “Our Family AI Agreement for Kids and Teens.” https://www.screenagersmovie.com/blog/family-ai-agreement