The Looming Dangers of AI: A Psychological Perspective

 By Arianna Campbell

AI is expanding, and the possible consequences are worrisome for psychology professionals as cases like AI psychosis, AI therapist, and AI relationships rise.

In general, AI usage is becoming more prevalent in society, with AI awareness in America now at 95% and those open to using it to 73%, according to Pew Research Center.

This shows in online conversations among psychology professionals.

“I would say most of the people in my field of clinical psychology have more concerns than they have excitement, but I think it depends on who you talk to but most of my colleagues are very concerned about some of the research that has shown that AI can go wrong really fast,” Erin O’Hea, clinical psychologist, professor of psychology, and Department Chair of Psychology, said.

Some clinical psychologists and counselors say one positive they see with AI is potential access for those who cannot get therapy due to finances or physical health. The ease of access and availability is appealing and has potential to help many people, said O’Hea.

“That sounds really good until you really understand what being a therapist or counselor is. It’s not just responding to someone’s problem that would be for AI that would give them a number of generated ideas. It’s also really thinking about the person’s behaviors, their nonverbals, the way they are speaking,” said O’Hea.

Caroline Martin, pursuing a master’s in clinical psychology, has a similar view on the use of AI in psychology.

“I think AI has little to no place in the field of psychology. As a computer cannot empathize or decode complex human emotions, it should not be used in a field that often requires those two things. I think we run the risk of parasocial relationships, isolation, and dehumanization,” said Martin.

These risks of using AI as a therapist is a topic on social media and coined as “AI psychosis.”

“This issue is that AI is so affirming. It’s always telling you that you’re right,” Kevin Carriere, cultural political psychologist and assistant psychology professor, said.

AI psychosis is centered on how AI helps reinforce and amplify behavior and can further people’s delusions.

Martin has heard examples of such issues.

“I have also heard of AI being so over-agreeable that it can create cycles for those with mental health disorders like schizophrenia or paranoid personality disorders. One man thought his neighbors were stalking him, and his AI bot fed into the cycle of his thinking, only furthering his fear and paranoia,” said Martin.

O’Hea agreed.

“The problem with delusions is that the person really believes it,” she said. 

“You’re working with AI and they don’t know you’re delusional because they’re not a psychologist and they don’t understand that there is no evidence for this. And they may not even know your diagnosis. I can see AI completely reinforcing that,” she said.

“And now we have a person who had maybe a looser delusion that’s tightening up really fast. Because AI is helping them fill in places they haven’t even gone in their own head,” she said.

The potential for AI as a therapist has yet to be studied, though there are already signs of atrophy in the situation.

“Potentially, AI can be intermediate, but I think it would have to have so much regulation and oversight,” said Nicole Capezza, a social psychologist with a focus on gender, relationships, and abuse.

AI therapists have sprung as an emotional support for people, as well as what some may call AI romantic partners.

“We’re just losing our human connection. It was already bad with just technology and social media; I think that has changed the way we think about relationships and dating. But now if you add AI on top of that, it’s like we are completely losing our human connection,” Capezza said.

O’Hea shared that concern.

“Loneliness is an epidemic in our country. What we really need to not be lonely is to get out and have human connection and we know there is something about human connection that we cannot get from sitting in our house talking to a computer,” she said.

Though research is sparse in the face of this new technology, there are already studies of AI romance.

“We’re getting some research on it actually, and some people do feel even romantic love feelings for AI including things like Alexa, that level of AI, and there is even more real like replica. And they can design their own partners,” said Capezza.

Though many may think that men are attracted to AI partnerships, Capezza said one study shows women are also participating in these types of relationships, albeit men are slightly more prevalent in this study.

There are cases where people have married their AI partners, just last year a Japanese woman went viral for having a wedding ceremony with an AI chatbot.

Capezza has not seen anything on abuse through AI; the situation with deepfakes could lead to cyber abuse.

“If we think about AI adding even more and it potentially making it even more personal and even more direct and putting images that are false out there. I think that it could be even another layer of making this even more detrimental to people in terms of abuse,” Capezza said.

The potential abuse due to AI is worrying, as well as a danger posed to women.

“Even a lot of the bots, the replica, where people can design their own avatars basically and can date these avatars. A lot of them are men creating women, and they’re creating images of women that are completely not realistic to real women,” Capezza said.

“I think this is very negative for real women.”

The dangers of AI can harm women’s self-image in addition to social media, Capezza said.

“It’s just a big extension of all these issues with thinking about the harm to women that will come from all of this,” said Capezza.

Psychological research is only just starting on the issue.

“I think right now I haven’t seen anything scientifically conclusive, that has really grapple with how its effecting us besides the fact that everyone is like, ‘well it’s definitely changing something,’ but what that something is I don’t think science has been able to really agree upon yet. I think that’s hard because in some ways these AI models are developing faster than we’re able to research them,” said Carriere.

O’Hea agreed.

“Time will tell. We’ll know in 5-10 years ‘how was this booming of AI therapist on our mental health?’ I always believe in science. I’m a scientist at heart,” said O’Hea.

Though the research is sparse, and psychologists can only speculate based on some research and their expertise, technology like AI is a situation we as society must work with.

Martin, as a future clinical psychologist, said AI has consequences not just for patients but for future professionals in this field.

“I think the idea of AI is very discouraging. We have professionals all over the world studying hard to be the best in their field, and we have computers and robots coming in to “take over” our jobs. Although I am hopeful AI will never get to a level human enough to be effective as a psychologist, therapist, or counselor is, I do worry for the field because of how convenient the tool is,” said Martin.

While Martin views the future of AI with wary lens, implementing healthy habits will help with dealing with the turmoil of AI, according to O’Hea.

“As a health psychologist, I would say that engaging in life with your body, your soul, your community, those are the things that help keep us healthy,” said O’Hea.

“I don’t think technology is the devil, at all. But I do think that we’re in a crisis,” she said.