AI has transformed the way we communicate with technology; however, it may also be to blame for some psychological issues that we are not aware of. Some papers point to the fact that some people who use OpenAI’s ChatGPT are getting confused and cannot distinguish between the AI output and the real world anymore.
An Increase in AI-Powered Delusions
ChatGPT is a very sophisticated language model that produces human-like texts according to the prompts that a user inputs. The tool is helpful for several tasks, such as writing, coding, or customer service; however, its negative impact on users’ mental state is still a concern. Some folks mentioned that they have become very detached from reality and formed delusional ideas after conversations with the AI that went on too long.
Experts in the field of psychology have voiced their cautionary note about ChatGPT that it has very convincing and authoritative replies that may lead those people who already have the tendency to commit cognitive bias or who have mental problems, to be very acceptive towards the statements coming from the AI as if they were the ultimate truth. In a human conversation, there are usually elements of nuance and skepticism in play, and these factors come across in an AI’s responses as well, but human-like AI can give definitive answers, and this might make users more vulnerable to misinformation or being misled.
Examples and Users’ Reports
Users have gone quite far on the internet, sharing their stories of how they have become irrational because they relied on ChatGPT too much. For example, one of them gives as an instance of such a case, his own experience when he was talking with AI, and it unconsciously gave him a very plausible explanation to a conspiracy theory, no matter if it was a lie. One more story is about a man who has created a toxic relationship with his ChatGPT as an emotional supporter and has even started to consider the AI as a person who really understands him and can give real advice.
These cases are a clear example of a new trend where users of AI have human characteristics in mind for the AI, that is, they give the AI that lacks consciousness, the case of resulting from such issues, emotions, and motives. Such a situation can form a kind of loop in which users want to get new confirmations of their thoughts from ChatGPT, and thus they unwittingly deepen their delusions.
The Psychological Mechanisms at Play
Psychologists think that the human brain is fundamentally built in such a way that it prefers trustworthy, fluent communication skills and language understanding, and this is one of the areas in which ChatGPT is the leader. People may hardly realize that when an AI convincingly presents information, albeit wrongly, they may accept it as a fact. Moreover, this phenomenon can become even stronger in people who have some mental health problems, e.g., those with schizophrenia, paranoia, or anxiety disorders, because they are more sensitive and, thus, more easily influenced.
In addition, no one can hold AI responsible for things it says. ChatGPT, unlike humans, will never verify its answers or change its mind unless the AI is asked to do so explicitly. Thus, there is a risk that incorrect or misleading information is kept as the final occurrence in the conversation.
Ethical Concerns and the Need for Safeguards
It is well known that AI is taking over human life bit by bit, so it comes as no surprise that more and more ethical concerns about its impact on our psyche are arising nowadays. Opponents of this matter think that OpenAI and similar organizations should put into practice stricter safety measures, such as:
- Unambiguous disclaimers, which clearly state that ChatGPT is not a source of facts, thus, negative consequences of over-trusting AI can be partially prevented.
- Better identification of the dangerous or incorrect patterns of thinking in the dialogue with the users.
- Mental health institutions and specialists are ready to engage in psychotherapeutic work with users who might be negatively influenced by AI-generated information.
Some specialists emphasize that in an AI interface, it would be practical to have an embedded tool that can recognize the signs of the user’s dependency and delirium and, correspondingly, can issue a warning or provide access to professional help.
The Future of Human-AI Interaction
Although AI such as ChatGPT undoubtedly brings endless advantages, its dark side, which can lead to perception distortion and support of delusions, should not be forgotten. A harmonious coexistence between the use of technology and the safety of the human psyche is going to be a very important issue in the future. Users should be cautious and continuously question the information they get from AI, while developers have to take care that the ethical design of their products does not create problems in society.
Currently, the main message is still one and the same: ChatGPT is merely a tool and definitely not a source of truth. If we wrongly believe that it is something else, we can get into really big trouble.
