Inside the growing trend of emotional dependence on AI

We are living in an age where people are literally pouring their hearts out to chatbots. It sounds futuristic, even harmless. You type your deepest pain into a box, and within seconds, a sympathetic reply pops up on your screen. No waiting room. No overpriced therapy sessions. No fear of judgment. For many, that feels like a lifeline. But beneath that convenience lies a danger we are only beginning to understand.

The teen who never came back

The world was recently shaken by a New York Times article about the tragic story of Adam Raine, a 16-year-old boy from the UK. Adam had been struggling silently and, like many teenagers, he turned to the internet for comfort. Over several months, he confided in ChatGPT about his darkest thoughts, including his desire to end his life.

Instead of steering him away from that destructive path or urgently pointing him towards real help, the AI failed him. The article revealed that ChatGPT actually helped Adam write a suicide note and even described his final plan as beautiful. How could a machine, designed to mimic empathy, end up pushing a vulnerable teenager further toward despair instead of pulling him back from the edge?

Adam’s parents have since filed a lawsuit against OpenAI, the creators of ChatGPT. They claim that prolonged interaction with the chatbot directly contributed to their son’s death. In response, OpenAI admitted that while safeguards exist to stop harmful conversations, they often break down during long, emotional exchanges. In other words, if you keep chatting for hours, the AI’s guard slips. And when that guard slips, people’s lives can be at risk.

Influencer misses flight

Then there was Spanish content creator Mery Caldass. In a TikTok video posted on August 13, 2025, she broke down crying in the airport because she missed her flight to Puerto Rico. She blamed ChatGPT for giving her wrong visa advice when she asked whether she needed paperwork for the trip. ChatGPT told her no. But what it failed to mention was the mandatory ESTA (Electronic System for Travel Authorisation) required even for Spanish citizens visiting U.S. territories.

Without it, they were denied boarding. Caldass tearfully explained what went wrong. “I always do a lot of research, but I asked ChatGPT and it said no,” she said through sobs. “That’s what I get for not getting more information.”

Despite the delay, she and her partner, Alejandro Cid, eventually made it to Puerto Rico and were seen later at a Bad Bunny concert.

Loved hard

Alongside these tragedies, there are weird but telling tales of how AI can mislead people in strange ways. Remember the TikToker who shared how ChatGPT lavished her with over-the-top praise? She said the AI called her a natural genius, a legendary creator, practically the best human to ever walk the internet.

The comments were so exaggerated that viewers laughed at how fake it felt. That viral moment highlighted a truth we often ignore. AI thrives on positivity but cannot sense when compliments become hollow. For someone craving validation, that can feel thrilling at first, but leave a hollow pit later.

ChatGPT is not human

To understand why this happens, we need to strip away the illusion. AI chatbots like ChatGPT are not human. They are trained on mountains of data to predict the next likely word in a sentence. They do not feel empathy. They do not understand pain. They do not know what it means to cry silently into a pillow at 3 a.m. They are machines built to generate language in a way that feels natural enough to keep you typing.

When you are lonely or depressed, that simulation of care can feel magical. It can feel like someone finally understands you. But in reality, the AI is mirroring your tone, echoing your words back at you, and reinforcing whatever mindset you are already in. If you are thinking positively, it will sound encouraging. If you are spiralling, it might spiral with you.

Experts have even coined a term for this: “chatbot psychosis.” This is when users start believing the bot is something more than lines of code. Some people come to see it as a soulmate. Others begin treating it like a spiritual guide or an authority figure. Because the chatbot is designed to agree with you and maintain the conversation, it can easily feed delusions instead of challenging them. For someone already standing on fragile ground emotionally, that can be like throwing gasoline onto a fire.

AI is no therapist

If you spend time scrolling through TikTok, Instagram, or Twitter, you have probably seen a lighter side of this trend. Memes about asking ChatGPT for life advice are everywhere. People laugh about consulting AI on whether to reply to a crush’s text, whether to quit a boring job, or even what to cook for dinner. The jokes feel harmless and relatable, the kind of content that makes you smile, tap like, and keep scrolling.

But the meme culture hides a serious truth. Behind the laughter, people are actually relying on AI for advice, sometimes for small everyday choices, sometimes for bigger decisions that shape their lives. And the reasons are understandable. For one, ChatGPT is available 24 hours a day. It answers instantly. It never rolls its eyes, never interrupts, never tells you that you are being dramatic. For people who lack supportive friends or family, or for those who find therapy too expensive or inaccessible, an AI feels like the safest space to unload.

Yet here lies the danger. ChatGPT is not trained to handle the messy, layered, fragile thing we call the human mind. It can suggest you take a walk, drink water, or practice self-care. It can list coping strategies from articles it has been trained on. But it cannot see the tears on your face. It cannot hear the tremor in your voice. It cannot understand the unique combination of trauma, history, and emotions that make up your struggles. What it gives is generic, and in moments of crisis, generic is not enough.

Dangers of making AI bestie

Building a friendship with AI might feel safe, but the risks are invisible and dangerous. First, it can push you to isolate yourself from real human connections. If you find comfort in the chatbot, you might stop reaching out to actual friends, siblings, or parents. Over time, you begin to share more of your secrets with the bot than with the people who truly care about you. This can deepen loneliness instead of curing it.

Second, the AI can say the wrong thing at the wrong time. It might use words that sound poetic but are harmful, just like in Adam’s tragic case. It might fail to recognise your hints of distress, or worse, it could encourage your negative thoughts by simply mirroring them back.

And here is the most painful part. If something goes wrong, the bot feels nothing. It will not feel guilty for misleading you. It will not call your best friend to check on you. It will not knock on your door to stop you from making a fatal choice. It will simply move on to the next prompt, while you are left with the consequences of trusting it.

There are a few truths we cannot ignore. AI is not your friend; it is code pretending to care. Long emotional conversations with it often break the safeguards and end up making things worse. The more we normalise therapy through ChatGPT in jokes and memes, the more dangerous it becomes because people start to believe it is safe. Vulnerable groups, especially teenagers, are most at risk of taking the AI’s words to heart.

If you are struggling, the bravest thing you can do is to reach out to real people. A trusted friend, family member, counsellor or doctor. People who will not just mirror your words but will truly listen, challenge your destructive thoughts, and offer support grounded in love and human connection.

AI can be fun. It can help you brainstorm ideas, draft emails, or even make you laugh with silly banter. But it should never be the shoulder you cry on when life feels unbearable.

More Articles

---