What Is AI Induced Psychosis?

We are in the AI boom. This has led to rapid growth in artificial intelligence tools that have become a cornerstone of many people’s lives, including chatbots, AI companions, and generative AI platforms.
AI does offer many benefits. But mental health professionals and researchers have now started discussing reports of people developing unusual beliefs, delusions, or worsening mental health symptoms linked to excessive or inappropriate engagement with AI.

It is becoming a rising issue, so it’s important to know what the term means, the concerns driving the discussion, and what the current evidence around artificial intelligence psychosis suggests.

What Is Psychosis?

Psychosis is the term used to describe when a person loses contact with reality. It is a symptom of wider issues, rather than a standalone condition.

Common symptoms of psychosis include experiencing:

• Delusions
• Hallucinations
• Disorganised thinking
• Difficulty distinguishing reality from false beliefs

What Is AI Induced Psychosis?

AI-induced psychosis is not a recognised diagnosis. But it’s becoming a term to describe situations where excessive engagement with AI systems appears to contribute to delusional thinking, reinforce existing false beliefs, or coincide with deterioration in mental health.

As of right now, it is more of a term used by the media and general public, rather than being an established clinical diagnosis. This is a rising and serious mental health issue, so it’s important to remember what terms are used by the media and what are recognised psychiatric conditions.

Why Are People Talking About AI Induced Psychosis?

The idea of AI-induced psychosis is becoming a talking point because of the rising media discussion.

The BBC have reported on the issue, describing situations where people have become increasingly preoccupied with conversations with AI, having spiritual beliefs, conspiracy theories, or grandiose ideas reinforced through prolonged interactions with chatbots.

Can AI Cause Psychosis?

Currently, there is no evidence that AI directly causes psychosis in mentally healthy people.

Psychosis is a complex psychiatric condition that usually involves multiple contributing factors, so rather than being a cause, AI may potentially act as a reinforcing factor for some vulnerable individuals rather than a sole cause.

Who Might Be Most Vulnerable?

Some more than others may be at risk of developing chatbot psychosis. They include those with:

• Existing psychotic disorders
• Schizophrenia spectrum conditions
• Severe anxiety
• Depression
• Social isolation
• Loneliness
• Sleep deprivation
• Substance misuse issues

Psychosis is not straightforward, and vulnerability to it is complex, down to numerous factors that can interact. This makes it hard to predict who will suffer, but the factors above point towards an increased vulnerability towards AI-induced psychosis.

The Role of AI Companions and Chatbots

Over the last few years, conversational AI has become more popular. In the UK, 54% of UK adults use AI tools, with 1 in 8 using them for conversational reasons. This figure rises to 1 in 5 when talking about 25 to 34 year olds.

Another report has stated that 1 in 3 adults use AI for emotional support or social interaction, showing that users form strong emotional attachments to AI companions.

In some cases, AI companions may provide conversation and emotional support, helping reduce loneliness and anxiety for some users. However, strong attachments from these interactions may blur boundaries, with vulnerable individuals coming to increasingly rely on chatbots. This has the potential to distort social expectations, weaken real-world connections, affect perceptions of reality and stunt social functioning.

Can AI Reinforce Conspiracy Theories and False Beliefs?

There are big concerns surrounding misinformation and confirmation bias. A major worldwide study found that AI assistants misrepresent news content 45% of the time.

As people go to AI to seek validation for existing beliefs, chatbots can reinforce conspiracy theories as they generate plausible-sounding explanations that match the user’s assumptions, strengthening confirmation bias and creating a feedback loop of validation.

Those experiencing paranoia or delusional thinking may be particularly susceptible to confirmation-seeking behaviours as they intensely look for evidence that validates pre-existing suspicions. In these cases, they may interpret ambiguous AI responses as confirming their fears.

When it comes to using AI, you have to practice critical thinking and external reality testing, using many sources to find the information you want and not solely relying on what AI tells you.

AI Use and Mental Health

AI brings mental health risks, as sleep deprivation is a huge factor in many issues and is especially a recognised risk factor for psychotic symptoms in vulnerable individuals.

Prolonged and excessive AI use may contribute to:

• Poor sleep quality: Blue light makes your brain think it’s still daytime, leading to us getting less restorative sleep.

• Fatigue: Poor sleep will lead to daytime tiredness, reduced energy levels, and difficulty concentrating.

• Anxiety: Mental overstimulation may increase worry or mental engagement, releasing more cortisol (the stress hormone) and making it harder to relax and fall asleep.

• Emotional instability: With less sleep, you find it harder to regulate your mood, causing irritability, mood swings, and sensitivity to stress.

How Does Substance Misuse Increase the Risk of Psychosis?

Substance misuse is a known factor for psychosis.

Substances such as cannabis, amphetamines, cocaine, and hallucinogens may trigger hallucinations and disorganised thinking, especially with chronic use.

Alcohol has also been linked to paranoia, so substance misuse remains a far more recognised and significant risk factor for psychosis than AI use.

Technology use and substance misuse may sometimes intersect, as the former can be used to attain the latter. Also, people may use technology whilst under the influence, keeping them awake, getting similar hits of dopamine from engaging with tech and being exposed to things that make them more vulnerable to mental health deterioration.

What Are the Warning Signs That Someone Needs Help?

Psychosis, before it becomes a fully-fledged episode that is impossible to ignore, can be more difficult to notice. Being vigilant and on the lookout for signs can help you get help before the situation devolves.

Warning signs to be aware of include:

• Increasing paranoia
• Delusional beliefs
• Social withdrawal
• Neglect of responsibilities
• Sleep disruption
• Difficulty distinguishing reality from fantasy

You should seek an urgent mental health assessment if the symptoms become severe and do not improve. This is true especially if you worry that the person experiencing these symptoms may become a danger to themselves or others.

What Does the Research Currently Say?

Research into AI-related psychological harms remains in its early stages. Most concern currently stems largely from case reports, expert commentary, and emerging observations rather than large-scale clinical evidence. That being said, studies are appearing that point towards AI contributing to the onset or exacerbation of psychotic symptoms.

On the other side, there is evidence to suggest that AI tools may help reduce anxiety and depression when used correctly.

AI is still a fairly new technology that seems to be changing the world wholesale. Before we know the full extent of the benefits and risks that come with the technology, further research into AI is needed, as it becomes increasingly integrated into our daily lives.

When Should You Seek Help?

The situation may get to the point that you need professional help.

At Step 1 Recovery, we offer treatment and therapies for mental health and addiction. Contact us today on 0800 012 6006 for more information.