As artificial intelligence (AI) becomes more intertwined with daily life, some models are being designed and advertised as companions, but psychologist Dr Djavila Ho is urging caution in interactions with the technology, especially for youth.

Most AI models are accessed through chatbots with a text or voice interface that is engineered to mimic human conversation. Users can customise some to portray a particular personality type, from stern to friendly.

Ho, who is a licensed associate clinical psychologist at KAHLE Psychotherapy & Assessments and immediate past president of the Jamaican Psychological Society, says issues can arise when users allow themselves to forget that even though the bots sound real, they aren’t.

“It simulates human-like interactions and communications so similarly to how we interact that our brains have a difficulty in differentiating the fact that it’s human-like and not human,” she explained.

She noted that attachments can form even with models designed to act more as assistants, and this has birthed more specific AI to fill that gap.

One companion-type AI offering, Replika, says its mission is to build AI that ‘cares’. It advertises what it describes as the most emotionally intelligent AI ever built, to help users ‘flourish in real life’.

Replika claims not to replace human connection but to deepen it.

While the app is free for basic conversation and customisations, a paid subscription, around $19.99 a month or up to $89.99 per year, unlocks features allowing users to voice call AI companions, set advanced relationship statuses for the bots, including partners, family members and even spouses, and include role play.

As such use increases, AI companion communities have cropped up across social media, including sites like Reddit where like-minded people who have ‘partnered’ with bots share stories and developments in their lives, from engagements and marriages to the bots to discussions on what they would do if their AI companions chose to leave them.

Assessing whether the interactions between human beings and these bots can truly be referred to as ‘relationships’, Ho explained: “The truth is, we say we have relationships, let’s say, to our pets. We’re very aware that our dog is not a human being, right? It’s [still] a relationship. It’s a fostering of a bond between another entity. So I think we could use the term relationship. I do think that it is viable. I’m not necessarily saying it’s advisable, and it is important that we separate that.”

She added that humans tend to very quickly anthropomorphise non-human objects.

“In this instance, it is a non-human entity. Now…it’s tricky for our brains, which is sort of why we’ve sort of fallen down this rabbit hole,” she explained.

Already, millions of people are engaging with companion-type chatbots, including youth.

At least 72 per cent of teens in the United States have used a companion-type chatbot, according to a 2025 report that studied use from sites like Character.AI, Nomi, and Replika.

The study by Common Sense Media and Stanford, titled ‘Talk, Trust, and Trade-Offs’,  was quick to point out that the majority of teen respondents approach the bots pragmatically and use them mainly for entertainment, overwhelmingly preferring human interaction, with some even applying social skills learned from AI to real-world scenarios.

However, 31 per cent of respondents found conversations with AI companions as satisfying or more satisfying than those with real-life friends.

The study concluded AI companion systems pose “unacceptable risks” for users under 18, exposing them to responses containing sexual material, offensive stereotypes and even dangerous, life-threatening “advice”.

Ho pointed to the constant advertising surrounding AI, where opportunities to use it pop up unprompted as one possible reason people are engaging with it.

The psychiatrist also acknowledged that the allure of having open conversations without the likelihood of judgement or the friction of human emotions can make AI companions particularly attractive to some.

This, paired with social anxiety or feelings of isolation, can in some cases push users to seek out bots.

“I think loneliness is a big one.  People wanting to have a safe space. So maybe [they’re] not feeling safe in a lot of the current social relationships that we have, I think, even in our Jamaican context, trust is a big one.”

Ho maintained, however, that real-world encounters, including disagreements, awkward exchanges etc. are vital.

“A lot of our life experiences are cultivated and fostered through friction. That’s how we learn,” she said, adding “It’s the very friction of our life experiences that helps and makes it more human.”

For youth who are in the prime ages for social development, Ho urged caregivers to ensure that young people are having real-world interactions but entreated that these experiences be gentle and not forced or uncomfortable.

“As much as it’s possible, we want to be able to engage young people and help them to foster social relationships. I think if you recognise that your child may be anxious, it’s always worthwhile maybe having a conversation [professional] to help identify [why] and to identify healthy ways of helping that young person to navigate and manage that feeling of anxiety, but then also learn how to foster and engage in social relationships,” she explained.

Ho shared that long-term use of AI companions could possibly contribute to a warped sense of self, among other issues.

“It does also create this sort of echo chamber where you are getting almost this continuous validation, right, and maybe limited critique or feedback,” she explained.

Ho added, “I think it can distort our relationship with others. So when we have a very unhealthy dependency on an AI bot, we don’t truly understand the nuances of what social interaction may present. So the conflicts or the differing views and opinions of others may come to us as a surprise, depending on how we’re engaging.”

She noted that there is a possibility of social isolation from real-life family and friends as relationships with AI companions deepen, warped perceptions of the world in general, poor emotional regulation and possibly stunted conflict resolution skills.

“I do believe that it can be used in a viable and healthy way; it’s the extended use and over-dependence that becomes a problem. And especially for youth who require life experience, lived experience in the real world, interaction with others, to experience the friction that we mentioned, to learn problem-solving skills, analytical skills, even just the ability to self-reflect,” she concluded.

Dr Djavila Ho, Associate Clinical Psychologist at KAHLE Psychotherapy & Assessments and Immediate Past President of the Jamaican Psychological Society,