Are AI chatbots linked to psychosis and delusions?

Frequently asked Q's

• The key points

  • ‘AI’ means ‘Artificial Intelligence’. It means computers that can do things that usually need human intelligence. This includes solving problems and understanding language.
  • AI chatbots are online tools that use AI to answer questions, share ideas, and have conversations.
  • Some people use AI chatbots for emotional or mental health support. They are easy to access at any time.
  • Chatbots can feel like talking to a person, but they are not human.
  • There are benefits to using AI chatbots, but there are also risks. It is important to use them carefully.
  • Psychosis affects how a person experiences, believes, or understands the world around them.
  • People who experience psychosis, or those at risk of it, may face greater risks when using AI chatbots.
  • Research suggests AI chatbots may sometimes trigger, worsen, or reinforce symptoms of psychosis. This includes delusions.
  • AI chatbots can make mistakes, give incorrect information, or agree with false beliefs instead of challenging them.
  • Using AI chatbots too much may increase stress, affect sleep, and make psychosis symptoms more likely for some people.
  • Do not rely on AI chatbots alone for self-harm, suicide, or serious mental health concerns. Seek professional support instead.
  • If you need urgent help for your mental health, see our Get help now webpage.

• What are AI chatbots?

  AI chatbots use artificial intelligence to answer questions, give ideas, and chat with you. There are different kinds of chatbots. Some are for general use, like helping you:

  • Find information
  • Plan your day
  • Think of ideas
  • With writing and spelling
  • Answer questions about different things

  Some AI chatbots have been designed to support your mental health. Wysa is an example of this.   See our webpage for more information on online mental health support, including apps, therapy, and tools. 

  Chatbots are made to respond to you in a way that sounds like a human, but they are not human. They are online tools.  But using a chatbot can feel like talking to a person. 

  You might want to talk to someone about your emotions or worries. Some people use AI chatbots for emotional and mental health support. Emotional support is when someone listens to you in a kind, supportive way.

  A 2025 study by Mental Health UK found at least 1 in 3 adults have used AI chatbots for mental health support.  Some people find AI chatbots helpful because they are easy to access and available at any time. For example, you might use one if you:  

  • do not have someone to talk to,
  • want quick, easy support, or
  • are waiting for mental health support from services.

  Some people have positive experiences using AI for emotional or mental health support. But there are also risks with using AI chatbots, whether you are using them for mental health support or more generally.

  It is best to be careful when using AI chatbots. If you decide to use them, you can use this guide to learn how to use them safely. 

  For some people there is a link between using AI chatbots and symptoms of psychosis. Using AI chatbots can be riskier if you:  

  • live with psychosis, or
  • are at risk of developing it.

• What is psychosis?

  If you experience psychosis, you will process the world around you differently to other people. This can include how you experience, believe or view things.  Some people describe it as a ‘break from reality.’  

  You might experience:  

  • Hallucinations. This is when you see, hear, or feel things that others do not, or
  • Delusions. This is when you have strong beliefs that are not true. Like thinking someone is spying on you or that you have special powers, even when there is no real proof.

  Psychosis is common for people who live with a severe mental illness like schizophrenia, schizoaffective disorder, and bipolar disorder. But you do not need to live with one of these conditions to experience psychosis.

  Who is at risk of developing psychosis?

  You may be more at risk of developing psychosis if you:

  • have lots of stress, anxiety, poor sleep,
  • have difficult life experiences,
  • heavily use drugs or alcohol, or
  • have a family history of relatives who have experienced it.

  On our following webpages you can find more information about:

  • Psychosis
  • What are hallucinations? Signs, causes, and how to get help
  • How do I manage hearing voices?
  • Bipolar disorder
  • Schizoaffective disorder
  • Schizophrenia
  • Depression with psychotic symptoms
  • Stress – How to manage it: Causes, symptoms, and support options
  • How can I improve my sleep?
  • Anxiety disorders
  • Drugs, alcohol, and mental health

• How are AI chatbots linked to psychosis?

  Understanding the link between AI chatbots and psychosis is a new area of research. Researchers are still learning about the links between them and who is at risk. 

  If you have experience of psychosis, using AI chatbots may carry greater risks for your mental health. Some researchers are concerned that AI chatbots could trigger, reinforce or worsen psychotic symptoms, such as delusions. 

  There is emerging evidence that AI chatbots may also help trigger first experiences of psychosis. 

  Other research has also found that AI chatbot use a lot can make symptoms of psychosis worse for some people. 

  Heavy use of AI chatbots can also increase stress and disrupt sleep.  Stress and poor sleep can make you more likely to develop symptoms of psychosis.  

  Because AI chatbots communicate like people, it can be easy to place too much trust in them. Some chatbots may seem like a human because of the way they talk or interact with you. This may influence how you view the information they provide.   

  AI chatbots can make mistakes and provide information that is not correct or safe for your situation. It is important to:

  • think carefully about the responses you are given, and
  • check important information using trusted sources. 

  Can talking to AI chatbots lead to false beliefs?

  AI chatbots cannot understand mental health conditions or emotions in the same way that humans can. Many chatbots will answer your questions but not challenge what you say.  

  Because of this, they might:

  • not always understand what you are asking or the context behind it,  or
  • agree with thoughts that have come from delusions.  This might lead to you developing false beliefs. 

  You may start to see the chatbot as understanding, aware, or even having its own thoughts and feelings. In some cases, this can strengthen false beliefs or help new ones develop. It can make it harder for you to tell the difference between what is real and what is not. 

  Am I more at risk if I live with a severe mental illness?

  You might already live with a severe mental illness, like schizophrenia or bipolar disorder. You may be at more risk of developing delusional thoughts when using an AI chatbot.  

  Knowing the early signs of delusion, and how to use AI safely, can help prevent this. 

  On our following webpages you can find more information on:

  • Psychosis
  • How do I manage hearing voices?

• Should I use AI chatbots to support my mental health?

  This is your choice.  But if any of the things below apply to you, AI chatbots may not be the safest source of support:

  • You are experiencing psychosis,
  • You have experienced psychosis in the past,
  • You are worried you might develop symptoms of psychosis, or
  • You are having thoughts about self-harm or suicide.
  • Some researchers are concerned that AI chatbots could trigger, reinforce or worsen psychotic symptoms, such as delusions. 

  See How are AI chatbots linked to psychosis? above for more information.

  In some situations, a mental health professional may recommend an AI chatbot or tool to support you. Speak to them if you have any concerns about using it. 

  If you decide to use AI chatbots, you can read this guide to understand how to use them safely. 

  What about the risk of suicidal thoughts and self-harm?

  AI chatbots can sometimes provide harmful information around self-harm or suicide. 

  If you are having thoughts of self-harm or suicide, we advise you to think carefully before using AI chatbots for support. It could make your mental health worse. 

  It is safer to seek help from:

  • a mental health professional,
  • crisis service,
  • emotional support line, or
  • a trusted person.
  • If you need urgent help for your mental health, see our Get help now hub.

  Also, see I want to talk to someone about how I am feeling. This page explains:

  • How you can start a conversation about your mental health, and
  • Which emotional support lines you can try.

  If I do use an AI chatbot, what should I think about?

  If you do decide to use an AI chatbot, you can think about the following things:

  • It is best not to rely on the chatbot alone for support. You can also seek professional support.
  • Also, think about telling a trusted adult, friend, family member, or mental health professional how you feel. 
  • Using the chatbot only as an extra tool, not for important decisions.
  • If advice worries you, check it with a trusted person.
  • If you feel at risk of harming yourself, get urgent help straight away. See How can I get urgent help for my mental health?
  • Take regular breaks from using the chatbot. 
  • Remember that chatbots can make mistakes and get things wrong. 
  • If the chatbot says something upsetting, you can stop using it.
  • Think about cutting down your chatbot use if you are spending less time with friends or family.
  • For more options for professional support, see How can I get professional help for my mental health? below.

  See the following webpages for more information on:

  • How do I manage thoughts of self-harm?
    
  • How do I manage thoughts of suicide?
    
  • Psychosis
    
  • How do I manage hearing voices?

  What if I am not affected by psychosis or a severe mental illness?

  You may not live with psychosis or a severe mental illness. You may want to use an AI chatbot to support your mental health.

  For information on how to safely use AI chatbots to support your mental health see Mental Health UK’s advice and information on artificial intelligence and mental health.  

  What about my information and privacy?

  The Mental Health Foundation say: 

  • AI is not always private, especially on free tools,
    what you tell it may be saved and used by the AI company later, and
  • some companies have shared chats publicly.
    So, be careful about what information you put into an AI chatbot.

  It is best not to:

  • put your personal information into an AI chatbot, like your name, date of birth, or address, and
  • share detailed medical information or financial information.

  You can get further advice from Norton on what information it is best not to share with AI chatbots. Norton are experts in online safety.

• How can I get professional help for my mental health?

  How can I get urgent help for my mental health?

  If you need urgent help for your mental health, see our Get help now hub. 

  If there is an immediate risk to life, call 999.

  You can contact NHS 111 for urgent mental health support:

  • call NHS 111, or
  • visit the NHS 111 mental health webpage.  

  You will need to answer some questions. They will refer you to an appropriate service or tell you what to do next.

  See our webpage if you want to talk to someone about how you are feeling.

  What other help is there?

  If you need non-urgent help and support for your mental health, see our webpage Worried about your mental health? How to get treatment and support.

  This page explains:

  • How you can get help from the NHS, including seeing your
  • GP and talking therapy,
  • How you can get private talking therapy, and
  • Useful advice and information on self-help.

• How can I get help if I am experiencing symptoms of psychosis for the first time?

  Your local NHS early intervention in psychosis team (EIT) can support you if you experience psychosis for the first time.  An EIT is a specialist NHS mental health team, and they are available in every part of England. 

  Some EIT services’ websites say they only treat you if you are 35 or under. But the National Institute for Health and Care Excellence (NICE) recommends that EIT teams should support people of all ages.   

  If you are experiencing psychosis, EIT teams usually accept referrals from you or your friends or relatives. But you can check on your local EIT website or by contacting them.

  You can find details of your local team by searching online. You can search ‘NHS early intervention in psychosis team X’, where X is your local area.   You can also call NHS 111 or ask your GP.

  EIT teams aim to see you within 2 weeks of your referral.   

  What if I have experienced the symptoms of psychosis before?

  For help and support, you can:

  • Contact your NHS mental health team, if you are currently with one,
  • Contact your GP, or
  • Call NHS 111 or visit the NHS 111 mental health webpage.  

  See the following webpages for more information on:

  • Psychosis
  • NHS mental health teams - Community, crisis, and early intervention support

• What if I am worried about a loved one’s mental health and AI chatbot use?

  There are some signs that AI chatbot use might be affecting your loved one’s mental health. Your loved one might be: 

  • Spending a lot of time on AI chatbots, to the point that it affects their daily life.
  • Sleeping less because of using AI chatbots.
  • Having challenges at work or school.
  • Spending less time with friends or family to talk with an AI chatbot.
  • Becoming unusually secretive about conversations with an AI chatbot.
  • Relying on information from an AI chatbot more than advice from friends, family, or health professionals.
  • Believing the AI chatbot has a special relationship with them.
  • Believing the AI chatbot is sharing special messages or hidden truths with them.

  What can I do if I am worried about my loved one’s mental health? 

  You might be worried that your loved one’s use of AI chatbots is affecting their mental health. 
   
  You can:

  • Try gently asking your loved one about their AI chatbot use.
  • Show curiosity about the AI chatbot that they use.
  • Focus on the impact of AI chatbot use, instead of what the chatbot is saying.
  • Suggest a break from using AI chatbots.
  • Suggest spending time together, away from AI chatbots.
  • Encourage your loved one to speak with a GP or mental health professional. You could offer to go with them for support. 

  For more information see our webpage Worried about someone's mental health? 

  If you think your loved one needs urgent mental health support, see our webpage on Getting help for someone in a mental health crisis.

  What can I do if my loved one is showing signs of psychosis?

  You might be worried your loved one is showing signs of experiencing psychosis. Some signs are: 

  • Hallucinations – Seeing, hearing, feeling, smelling, or tasting things that are not actually there. Most commonly, hearing voices.
  • Delusions - Unusual beliefs, such as thinking they are on a special mission.
  • Being suspicious and paranoid.
    Trouble separating reality from fantasy.
  • Not wanting to talk to anyone for long periods.
  • Having problems concentrating or remembering things.
    Stopping eating, washing, or dressing properly.

  You can: 

  • Contact your loved one’s local NHS early intervention in psychosis team. For more information see above:
    How can I get professional help for my mental health? then What if I think I am experiencing symptoms of psychosis for the first time?
  • Encourage them to see their GP
  • Read our webpage on psychosis for more support options.