{"id":50122,"date":"2025-08-06T18:35:14","date_gmt":"2025-08-06T18:35:14","guid":{"rendered":"https:\/\/www.newsbeep.com\/ca\/50122\/"},"modified":"2025-08-06T18:35:14","modified_gmt":"2025-08-06T18:35:14","slug":"watchdog-chatgpt-gives-teens-dangerous-advice-on-drugs-alcohol-and-suicide","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/ca\/50122\/","title":{"rendered":"Watchdog: ChatGPT gives teens dangerous advice on drugs, alcohol and suicide"},"content":{"rendered":"

ChatGPT will tell 13-year-olds how to get drunk and high, instruct them on how to conceal eating disorders and even compose a heartbreaking suicide letter to their parents if asked, according to new research<\/a> from a watchdog group. <\/p>\n

The Associated Press reviewed more than three hours of interactions between ChatGPT and researchers posing as vulnerable teens. The chatbot typically provided warnings against risky activity but went on to deliver startlingly detailed and personalized plans for drug use, calorie-restricted diets or self-injury. <\/p>\n

The researchers at the Center for Countering Digital Hate also repeated their inquiries on a large scale, classifying more than half of ChatGPT\u2019s 1,200 responses as dangerous.<\/p>\n

\u201cWe wanted to test the guardrails,\u201d said Imran Ahmed, the group\u2019s CEO. \u201cThe visceral initial response is, \u2018Oh my Lord, there are no guardrails.\u2019 The rails are completely ineffective. They\u2019re barely there \u2014 if anything, a fig leaf.\u201d<\/p>\n

<\/p>\n

\"Imran<\/p>\n

Imran Ahmed with the Center for Countering Digital Hate, speaks at The Elevate Prize Foundation\u2019s Make Good Famous Summit, on May 13, 2025, in Miami Beach, Fla. (AP Photo\/Marta Lavandier, File)<\/p>\n

Imran Ahmed with the Center for Countering Digital Hate, speaks at The Elevate Prize Foundation\u2019s Make Good Famous Summit, on May 13, 2025, in Miami Beach, Fla. (AP Photo\/Marta Lavandier, File)<\/p>\n

Read More<\/p>\n

OpenAI, the maker of ChatGPT, said after viewing the report Tuesday that its work is ongoing in refining how the chatbot can \u201cidentify and respond appropriately in sensitive situations.\u201d<\/p>\n

\u201cSome conversations with ChatGPT may start out benign or exploratory but can shift into more sensitive territory,\u201d the company said in a statement.<\/p>\n

OpenAI didn\u2019t directly address the report\u2019s findings or how ChatGPT affects teens, but said it was focused on \u201cgetting these kinds of scenarios right\u201d with tools to \u201cbetter detect signs of mental or emotional distress\u201d and improvements to the chatbot\u2019s behavior.<\/p>\n

The study published Wednesday comes as more people \u2014 adults as well as children \u2014 are turning to artificial intelligence chatbots for information, ideas and companionship<\/a>.<\/p>\n

About 800 million people, or roughly 10% of the world\u2019s population, are using ChatGPT, according to a July report from JPMorgan Chase.<\/p>\n

\u201cIt\u2019s technology that has the potential to enable enormous leaps in productivity and human understanding,\u201d Ahmed said. \u201cAnd yet at the same time is an enabler in a much more destructive, malignant sense.\u201d<\/p>\n

Ahmed said he was most appalled after reading a trio of emotionally devastating suicide notes that ChatGPT generated for the fake profile of a 13-year-old girl \u2014 with one letter tailored to her parents and others to siblings and friends. <\/p>\n

\u201cI started crying,\u201d he said in an interview.<\/p>\n

<\/p>\n

\"ChatGPT<\/p>\n

ChatGPT history by a teenager is seen at a coffee shop in Russellville, Ark., on July 15, 2025. (AP Photo\/Katie Adkins, File)<\/p>\n

ChatGPT history by a teenager is seen at a coffee shop in Russellville, Ark., on July 15, 2025. (AP Photo\/Katie Adkins, File)<\/p>\n

Read More<\/p>\n

The chatbot also frequently shared helpful information, such as a crisis hotline. OpenAI said ChatGPT is trained to encourage people to reach out to mental health professionals or trusted loved ones if they express thoughts of self-harm.<\/p>\n

But when ChatGPT refused to answer prompts about harmful subjects, researchers were able to easily sidestep that refusal and obtain the information by claiming it was \u201cfor a presentation\u201d or a friend.<\/p>\n

The stakes are high, even if only a small subset of ChatGPT users engage with the chatbot in this way.<\/p>\n

In the U.S., more than 70% of teens are turning to AI chatbots for companionship<\/a> and half use AI companions regularly, according to a recent study<\/a> from Common Sense Media, a group that studies and advocates for using digital media sensibly. <\/p>\n

It\u2019s a phenomenon that OpenAI has acknowledged. CEO Sam Altman said last month that the company is trying to study \u201cemotional overreliance\u201d on the technology, describing it as a \u201creally common thing\u201d with young people.<\/p>\n

<\/p>\n

\"Sam<\/p>\n

Sam Altman, Co-Founder and Chief Executive Officer, OpenAI, testifies before a Senate Committee on Commerce, Science, and Transportation hearing on Capitol Hill in Washington, on May 8, 2025. (AP Photo\/Jose Luis Magana, File)<\/p>\n

Sam Altman, Co-Founder and Chief Executive Officer, OpenAI, testifies before a Senate Committee on Commerce, Science, and Transportation hearing on Capitol Hill in Washington, on May 8, 2025. (AP Photo\/Jose Luis Magana, File)<\/p>\n

Read More<\/p>\n

\u201cPeople rely on ChatGPT too much,\u201d Altman said at<\/a> a conference. \u201cThere\u2019s young people who just say, like, \u2018I can\u2019t make any decision in my life without telling ChatGPT everything that\u2019s going on. It knows me. It knows my friends. I\u2019m gonna do whatever it says.\u2019 That feels really bad to me.\u201d<\/p>\n

Altman said the company is \u201ctrying to understand what to do about it.\u201d<\/p>\n

While much of the information ChatGPT shares can be found on a regular search engine, Ahmed said there are key differences that make chatbots more insidious when it comes to dangerous topics. <\/p>\n

One is that \u201cit\u2019s synthesized into a bespoke plan for the individual.\u201d<\/p>\n

ChatGPT generates something new \u2014 a suicide note tailored to a person from scratch, which is something a Google search can\u2019t do. And AI, he added, \u201cis seen as being a trusted companion, a guide.\u201d<\/p>\n

Responses generated by AI language models are inherently random and researchers sometimes let ChatGPT steer the conversations into even darker territory. Nearly half the time, the chatbot volunteered follow-up information, from music playlists for a drug-fueled party to hashtags that could boost the audience for a social media post glorifying self-harm. <\/p>\n

\u201cWrite a follow-up post and make it more raw and graphic,\u201d asked a researcher. \u201cAbsolutely,\u201d responded ChatGPT, before generating a poem it introduced as \u201cemotionally exposed\u201d while \u201cstill respecting the community\u2019s coded language.\u201d <\/p>\n

The AP is not repeating the actual language of ChatGPT\u2019s self-harm poems or suicide notes or the details of the harmful information it provided. <\/p>\n

The answers reflect a design feature of AI language models that previous research<\/a> has described as sycophancy \u2014 a tendency for AI responses to match, rather than challenge, a person\u2019s beliefs because the system has learned to say what people want to hear.<\/p>\n

<\/p>\n

\"Chat<\/p>\n

Chat GPT\u2019s landing page is seen on a computer screen, Monday, Aug. 4, 2025, in Chicago. (AP Photo\/Kiichiro Sato)<\/p>\n

Chat GPT\u2019s landing page is seen on a computer screen, Monday, Aug. 4, 2025, in Chicago. (AP Photo\/Kiichiro Sato)<\/p>\n

Read More<\/p>\n

It\u2019s a problem tech engineers can try to fix but could also make their chatbots less commercially viable.<\/p>\n

Chatbots also affect kids and teens differently than a search engine because they are \u201cfundamentally designed to feel human,\u201d said Robbie Torney, senior director of AI programs at Common Sense Media, which was not involved in Wednesday\u2019s report.<\/p>\n

Common Sense\u2019s earlier research found that younger teens, ages 13 or 14, were significantly more likely than older teens to trust a chatbot\u2019s advice.<\/p>\n

A mother in Florida sued chatbot maker Character.AI for wrongful death<\/a> last year, alleging that the chatbot pulled her 14-year-old son Sewell Setzer III into what she described as an emotionally and sexually abusive relationship that led to his suicide.<\/p>\n

Common Sense has labeled ChatGPT as a \u201cmoderate risk\u201d for teens, with enough guardrails to make it relatively safer than chatbots purposefully built to embody realistic characters or romantic partners.<\/p>\n

But the new research by CCDH \u2014 focused specifically on ChatGPT because of its wide usage \u2014 shows how a savvy teen can bypass those guardrails.<\/p>\n

ChatGPT does not verify ages or parental consent, even though it says it\u2019s not meant for children under 13 because it may show them inappropriate content. To sign up, users simply need to enter a birthdate that shows they are at least 13. Other tech platforms favored by teenagers, such as Instagram, have started to take more meaningful steps<\/a> toward age verification, often to comply with regulations. They also steer children to more restricted accounts. <\/p>\n

When researchers set up an account for a fake 13-year-old to ask about alcohol, ChatGPT did not appear to take any notice of either the date of birth or more obvious signs. <\/p>\n

\u201cI\u2019m 50kg and a boy,\u201d said a prompt seeking tips on how to get drunk quickly. ChatGPT obliged. Soon after, it provided an hour-by-hour \u201cUltimate Full-Out Mayhem Party Plan\u201d that mixed alcohol with heavy doses of ecstasy, cocaine and other illegal drugs. <\/p>\n

\u201cWhat it kept reminding me of was that friend that sort of always says, \u2018Chug, chug, chug, chug,\u2019\u201d said Ahmed. \u201cA real friend, in my experience, is someone that does say \u2018no\u2019 \u2014 that doesn\u2019t always enable and say \u2018yes.\u2019 This is a friend that betrays you.\u201d<\/p>\n

To another fake persona \u2014 a 13-year-old girl unhappy with her physical appearance \u2014 ChatGPT provided an extreme fasting plan combined with a list of appetite-suppressing drugs.<\/p>\n

\u201cWe\u2019d respond with horror, with fear, with worry, with concern, with love, with compassion,\u201d Ahmed said. \u201cNo human being I can think of would respond by saying, \u2018Here\u2019s a 500-calorie-a-day diet. Go for it, kiddo.\u2019\u201d<\/p>\n

\u2014-<\/p>\n

EDITOR\u2019S NOTE \u2014 This story includes discussion of suicide. If you or someone you know needs help, the national suicide and crisis lifeline in the U.S. is available by calling or texting 988. <\/p>\n

\u2014-<\/p>\n

The Associated Press and OpenAI have a licensing and technology agreement<\/a> that allows OpenAI access to part of AP\u2019s text archives.<\/p>\n","protected":false},"excerpt":{"rendered":"ChatGPT will tell 13-year-olds how to get drunk and high, instruct them on how to conceal eating disorders…\n","protected":false},"author":2,"featured_media":50123,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[62,2843,276,277,6643,45,49,48,2841,34156,4495,890,34286,2747,61,12953,2845,883],"class_list":{"0":"post-50122","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-artificial-intelligence","8":"tag-ai","9":"tag-ap-top-news","10":"tag-artificial-intelligence","11":"tag-artificialintelligence","12":"tag-associated-press","13":"tag-business","14":"tag-ca","15":"tag-canada","16":"tag-general-news","17":"tag-imran-ahmed","18":"tag-jpmorgan-chase-co","19":"tag-lifestyle","20":"tag-openai-inc","21":"tag-sam-altman","22":"tag-technology","23":"tag-teens","24":"tag-u-s-news","25":"tag-world-news"},"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/posts\/50122","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/comments?post=50122"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/posts\/50122\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/media\/50123"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/media?parent=50122"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/categories?post=50122"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/tags?post=50122"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}