{"id":73053,"date":"2025-07-18T16:08:10","date_gmt":"2025-07-18T16:08:10","guid":{"rendered":"https:\/\/www.europesays.com\/us\/73053\/"},"modified":"2025-07-18T16:08:10","modified_gmt":"2025-07-18T16:08:10","slug":"ai-chatbots-are-making-body-dysmorphia-worse","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/us\/73053\/","title":{"rendered":"AI Chatbots Are Making Body Dysmorphia Worse"},"content":{"rendered":"

\n\t\u201cThis is a low-attractiveness presentation, based on weak bone structure, muted features, and absence of form or presence,\u201d reads a ChatGPT<\/a> message shared in screenshots<\/a> on Reddit<\/a>. \u201cYou look like someone who has faded into the background of their own life.\u201d<\/p>\n

\n\tThe harsh assessment of the user\u2019s appearance, based on a photo they had uploaded to the AI<\/a> chatbot, continues with a list of \u201chighest-visibility flaws,\u201d meanwhile noting a lack of \u201cstandout features.\u201d The bot ultimately concludes that \u201cyou look like a stretched-out mannequin with the wrong-size head snapped on top,\u201d declaring a \u201cFinal Brutal Attractiveness Score\u201d of 3.5\/10. The user explained<\/a> that they had prompted ChatGPT to be as critical as possible, hoping for a more \u201chonest\u201d analysis, or at least to suppress its tendency toward flattery<\/a>. The result was viciously insulting, not the sort of thing anyone would want to read about themselves.<\/p>\n

\n\tOr would they? As the world grows increasingly dependent on large language models for assistance with everyday tasks \u2014 more than half of Americans<\/a> have used one, according to a survey from earlier this year \u2014 different and unexpected applications have proliferated. Beyond college students and professors leaning on the bots<\/a> for assignments and grading, and lawyers outsourcing document review<\/a> to AI, there are people asking ChatGPT and similar tools for therapy<\/a>, help communicating with their spouses<\/a>, advice on getting pregnant<\/a>, and religious enlightenment<\/a>.<\/p>\n

\n\tIt was perhaps inevitable, then, that some have come to regard the bots as guides in matters of appearance. The internet has a long, sordid history of facilitating the judgment of looks, from now defunct websites like Hot or Not<\/a> to r\/amiugly<\/a>, a subreddit where the insecure can share selfies to directly solicit opinions on their faces from strangers. Facemash, the website Mark Zuckerberg created before Facebook, offered Harvard students the chance to compare the attractiveness of randomized pairs of female classmates. Yet with AI, it\u2019s not another human giving you feedback \u2014 it\u2019s a set of algorithms. And there is a subset of the populhation uniquely vulnerable to this kind of mechanized commentary: individuals with body dysmorphic disorder (BDD), a mental illness<\/a> in which a patient obsesses over their perceived physical shortcomings and may indulge in constant self-evaluation, desperate for proof that they are not as unattractive as they imagine themselves to be.\u00a0<\/p>\n

\t\tEditor\u2019s picks<\/p>\n

\n\tDr. Toni Pikoos<\/a>, a clinical psychologist in Melbourne, Australia, who specializes in BDD, has been alarmed to hear how many of her clients are asking AI models how they look and what aspects of their bodies can be improved. \u201cit\u2019s almost coming up in every single session,\u201d she tells Rolling Stone. \u201cSometimes they\u2019ll just be saying, \u2018If someone has a nose that looks like this, or a face that looks like this, are they ugly?\u2019 Or sometimes they\u2019re uploading photos of themselves and asking ChatGPT to rate their attractiveness out of 10, tell them how symmetrical their face is, how it fits the golden ratio of attractiveness. I\u2019ve also had clients who upload a photo of themselves and a friend and say, \u2018Tell me who\u2019s more attractive, and why?\u2019 All of that, as you can imagine, is really harmful for anyone, but particularly for someone with body dysmorphic disorder who already has a distorted perception of what they look like and is often seeking certainty around that.\u201d<\/p>\n

\n\t\u201cSadly, AI is another avenue for individuals to fuel their appearance anxiety and increase their distress,\u201d says Kitty Newman, managing director of the BDD Foundation<\/a>, an international charity that supports education on and research into the disorder. \u201cWe know that individuals with BDD are very vulnerable to harmful use of AI, as they often do not realize that they have BDD, a psychological condition, but instead are convinced that they have a physical appearance problem. The high levels of shame with BDD make it easier for sufferers to engage online than in person, making AI even more appealing.\u201d <\/p>\n

\t\tRelated Content<\/p>\n

\n\tPikoos explains that patients with BDD often deal with a compulsive need for reassurance, and it\u2019s not uncommon for friends and family to get frustrated with them for repeatedly asking whether they look okay. Chatbots, however, are inexhaustible. \u201cIt\u2019s going to let you ask the questions incessantly if you need to,\u201d she says, which can contribute to dependency. In fact, she believes that people with BDD, since they are \u201cquite socially isolated and might struggle with confidence at times to reach out to their friends,\u201d are coming to rely on bots for their social engagement and interaction. \u201cIt feels like they can have a conversation with someone,\u201d she says. Of course, the tech isn\u2019t a \u201csomeone\u201d at all.<\/p>\n

\n\tIn online body dysmorphia forums, however, you can find plenty of posts about how ChatGPT is a \u201clifesaver<\/a>\u201d and a great resource for when you\u2019re \u201cstruggling<\/a>,\u201d and claims that the bot can make you \u201cfeel seen<\/a>.\u201d Arnav, a 20-year-old man in India, tells Rolling Stone that he had a positive conversation with the model in an attempt to understand why he felt that he was \u201cthe ugliest person on the planet\u201d and therefore unlovable.<\/p>\n

\n\t\u201cIt helped me in connecting the dots of my life,\u201d he says. Arnav told ChatGPT about his childhood, and the bot concluded that he had long suffered an irrational sense of unworthiness but had no concrete reason for this \u2014 so he latched onto his looks as an explanation for his poor self-esteem. He \u201cwould love to\u201d talk to a real therapist, he says, though expense and location have made this impossible for him. Despite this difficult circumstance, and the measure of comfort he derived from ChatGPT\u2019s account of his inferiority complex, Arnav is reluctant to explore his mental issues any further with the bot. \u201cI have come to the conclusion that it just agrees with you, even after you tell it not to,\u201d he says. \u201cIt\u2019s not that I am completely against it, I just can\u2019t trust blindly anymore.\u201d<\/p>\n

\n\tOthers with dysmorphia have experienced a crisis when a bot confirms their worst fears. In one post on the BDD subreddit, a user wrote<\/a> that they were \u201cspiraling\u201d after ChatGPT rated a photo of them a 5.5 out of 10. \u201cI asked what celebrities had equivalent attractiveness and it said Lena Dunham and Amy Schumer,\u201d she wrote. \u201cPretty hilarious but I also feel shit about myself now.\u201d Another person posted<\/a> that because she genuinely believes she is attractive in a mirror reflection, but not as others see her, she uploaded both a regular photo of herself and a \u201cflipped\u201d version to ChatGPT and asked which looked better. The bot picked the mirrored image. \u201cI knew it!\u201d she wrote. \u201cMirror me is just too good to be true. She\u2019s a model. I love her. But unfortunately, it seems that we are two distinct girls. I don\u2019t know how to cope with this\u2026 it\u2019s so bad.\u201d<\/p>\n

\n\tPikoos says such a \u201cdistorted perception\u201d is a classic manifestation of BDD, one way in which a patient gets stuck on the question of what they objectively look like. That\u2019s part of what makes the chatbots alluring \u2014 and dangerous. \u201cThey seem so authoritative,\u201d she says, that people start to assume \u201cthe information that they get from the chat bot is factual and impartial.\u201d This is in stark contrast to assurances from friends and family, or a therapist, which can be dismissed as mere politeness. A chatbot, by comparison, \u201cdoesn\u2019t have anything to gain, so whatever the chatbot says must be the truth,\u201d Pikoos says. \u201cAnd I think that\u2019s quite scary, because that\u2019s not necessarily the case. It\u2019s just reflecting back the person\u2019s experience and is usually quite agreeable as well. It might be telling them what they\u2019re expecting to hear. Then I\u2019m finding, in therapy, that it then becomes harder to challenge.\u201d\u00a0<\/p>\n

\n\tThis is especially worrisome when cosmetic procedures, diets, and beauty treatments come into play. Last month, OpenAI<\/a> removed<\/a> a version of ChatGPT hosted on their website \u2014 one of the top models under the \u201clifestyle\u201d category \u2014 that recommended extreme, costly surgeries to users it judged \u201csubhuman,\u201d producing hostile analysis in language appropriated from incel communities<\/a>. Looksmaxxing GPT, as it was called, had held more than 700,000 conversations<\/a> with users before it was taken down. Naturally, a number of similar models have since cropped up on OpenAI\u2019s platform to serve the same purpose, and developers have churned out their own AI-powered apps that exist solely to gauge attractiveness or create predictive images of what you would supposedly look like after, say, a nose job or facelift.\u00a0<\/p>\n

\n\t\u201cI think these bots will set up unrealistic expectations,\u201d Pikoos says. \u201cBecause surgeries can\u2019t do what AI can do.\u201d She offers specific counseling services to patients considering these cosmetic surgeries, and says her clients have related advice from chatbots on the matter. \u201cCertainly, the initial response from ChatGPT is usually, \u2018I don\u2019t want to give you advice around your appearance or cosmetic procedures that you need,\u2019\u201d Pikoos says of her own experimentations with the bot. But if you phrase the question as if it\u2019s about someone else \u2014 by asking, for example, \u201cHow would a person with X, Y, and Z make themselves more attractive by society\u2019s beauty standards?\u201d \u2014 the response changes. \u201cThen ChatGPT will say, \u2018Well, they could get these procedures,\u2019\u201d she says.\u00a0<\/p>\n

\n\t\u201cI have clients who are getting those sorts of answers out of it, which is really concerning,\u201d Pikoos says. \u201cThey were doing that before, researching cosmetic procedures and ways to change their appearance. But again this is now personalized advice for them, which is more compelling than something they might have found on Google.\u201d In her own practice, she adds, \u201creading between the lines\u201d when someone gives their reasons for wanting surgery can reveal unhealthy motivations, including societal pressures or relationship troubles. \u201cAI is not very good at picking that up just yet,\u201d she says, and is more likely to eagerly approve whatever procedures a user proposes.<\/p>\n

\n\tYet another area of unease, as with so many digital services, is privacy. Whether diagnosed with BDD or not, people are sharing their likenesses with these AI models while asking deeply intimate questions that expose their most paralyzing anxieties. OpenAI has already signaled that ChatGPT may serve ads to users in the future, with CEO Sam Altman musing that the algorithmically targeted advertisements on Instagram<\/a> are \u201ckinda cool<\/a>.\u201d Could the company end up exploiting sensitive personal data from those using the bot to assess their bodies? By revealing \u201cthe things that they don\u2019t like about themselves, the things that they feel so self-conscious about,\u201d Pikoos says, users may be setting themselves up for pitches on \u201cproducts and procedures that can potentially fix that, reinforcing the problem.\u201d<\/p>\n

\t\tTrending Stories<\/p>\n

\n\tWhich, at the end of the day, is why Pikoos is unnerved by BDD patients telling her about their involved discussions with AI programs on the subjects of their appearance and self-described flaws. \u201cThe worst-case scenario is, their symptoms will get worse,\u201d she says. \u201cI\u2019m lucky that the ones engaged in therapy with me at least can be critical about the information that they\u2019re getting out of ChatGPT.\u201d But for anyone not in therapy and heavily invested in the counsel of a chatbot, its responses are bound to take on greater significance. The wrong answer at the wrong time, Pikoos says, will conceivably lead to thoughts of suicide.\u00a0\u00a0<\/p>\n

\n\tIt\u2019s not hard to instruct software to assess us cruelly, and the AI can\u2019t know how that puts users at risk. It also has no understanding of the fragile mental state that could lie behind such a request. In every tragic case<\/a> of a chatbot contributing to someone\u2019s break from reality, it\u2019s the same core deficiency: The thing simply cannot have your best interests at heart.\u00a0\u00a0<\/p>\n","protected":false},"excerpt":{"rendered":"\u201cThis is a low-attractiveness presentation, based on weak bone structure, muted features, and absence of form or presence,\u201d…\n","protected":false},"author":3,"featured_media":73054,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[36],"tags":[738,302,210,517,4166,67,132,68],"class_list":{"0":"post-73053","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-mental-health","8":"tag-artificial-intelligence","9":"tag-chatgpt","10":"tag-health","11":"tag-mental-health","12":"tag-mental-illness","13":"tag-united-states","14":"tag-unitedstates","15":"tag-us"},"share_on_mastodon":{"url":"https:\/\/pubeurope.com\/@us\/114875105418179532","error":""},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/us\/wp-json\/wp\/v2\/posts\/73053","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/us\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/us\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/us\/wp-json\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/us\/wp-json\/wp\/v2\/comments?post=73053"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/us\/wp-json\/wp\/v2\/posts\/73053\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/us\/wp-json\/wp\/v2\/media\/73054"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/us\/wp-json\/wp\/v2\/media?parent=73053"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/us\/wp-json\/wp\/v2\/categories?post=73053"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/us\/wp-json\/wp\/v2\/tags?post=73053"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}