AI is now rated as more emotionally intelligent than humans (Ovsyannikova, 2025). It has the potential to change the way we form relationships, with a number of possible costs and benefits. It can help us interact more productively with dating apps, support us to date in a more healthy way, help us prepare and enhance our interpersonal skills, and it won’t be long before it can find us compatible people without dating apps.
Chatbots could help with the tedious parts of online dating for us
Many people use general chatbots such as Gemini, Claude and Chat GPT to get help with creating an effective dating profile, suggesting icebreakers and conversation starters for messaging, and asking common questions about dating. There are specific AIs to enhance your dating profile and suggest messages too. One man found his wife on Tinder using Chat GPT and AI bots on the Match.com dating apps will soon be available to support profiles and messaging. There are numerous risks here, especially for vulnerable people who could be even more easily manipulated into scams or abuse, and the difficulty of translating a great online conversation into real life. But there are some benefits too, especially for people who find it hard to message but are good in person, such as many in the neurodivergent community.
Rather than focusing resources on AI support, however, the dating apps could use AI more effectively to sort and match people. At the moment they claim to find you compatible partners through algorithms, but these often do not use the best research about what predicts satisfying, stable relationships. The current algorithms might be fine if you’re looking for something short-term or casual, but for people looking for a relationship, this is sub-optimal. Many daters report dating apps as worse than a few years ago. “Bumble is very glitchy, and on Hinge I once got shown people over 60 for a few days, and often I’m shown people that don’t live anywhere near me,” says Nick, 31. “Since I came back on the apps in 2024 after a break of 4 years, I’ve experienced so much more flakiness and ghosting. Guys seem to want to talk for ages but if you ask them to meet for a drink they ghost,” says Charlotte, 36. Bumble has just reduced its staff by 30%, reflecting the decreasing popularity of online dating – but will it improve its automated processes?
As I’ve explored elsewhere, there are other problems with the algorithms regarding bias, including the evidence that they perpetuate bias and discrimination if they are not designed and trained properly. Eventually, though, AI will be able to search the whole internet to find us matches without the use of apps, reducing the time and effort required to find a compatible partner.
AI can give us more personalised recommendations
AI-powered apps and platforms can already recommend events, activities, experiences, and products based on our personal preferences, potentially facilitating the formation of connections with others who share our interests. Now that we can choose whether a chatbot stores – and is informed by – previous conversations, we could allow the chatbot to make recommendations about our dating. (Of course, we should research and regularly review privacy issues here regarding data storage and sharing.) For example, the chatbot could suggest what to wear for a date (or or comment on your choice) or where to go, or recommend a suitable local singles event or activity for meeting new people.
Is this useful? In our journey of personal growth, research from coaching shows that unsolicited advice or recommendations are less likely to be effective than solutions we come up with for ourselves (Rechberg & Hovhannisyan, 2023). And it’s worth bearing in mind our autonomy and sense of agency (Kirk et al., 2025), checking out recommendations for ourselves and reflecting on whether they are aligned with our values. On the other hand, there may be benefits, such as giving us information about events we don’t have time to search for, reassurance about our sartorial judgment, or the push we need to get out there and socialise.
AI chatbots could be companions on the healthy dating journey
The use of AI is now for many part of daily life for administrative and creative tasks, and some enjoy emotional support and companionship, too (Fang et al., 2025). These could be useful for help with dating and relationships. When we catastrophise about how much we hate dating, our chatbot could remind us that of our last ten dates, we enjoyed five. When we get anxious before a date, the chatbot could suggest a meditation that calmed us down last time. When we are tempted to message an ex, the chatbot could remind us of what happened last time and suggest other more healthy ways to meet our emotional and sexual needs. (These kind of conversations would again require us to allow the chatbot to access our historical conversations with them, so it’s wise to regularly review the privacy policy of the AI provider.)
In relationships, such chatbots could remind us of anniversaries, give us requested feedback on our argument styles (maybe even after recording an audio of an actual argument) or suggest gentler ways to express our needs.
But there are dangers here of becoming reliant on the chatbot for reassurance, problem-solving, or even emotion regulation. And will we actually internalise learning if we don’t discover it for ourselves (Kolb, 2014)? Let’s look at the possible costs and benefits.
There are pros and cons of using chatbots for support
There are many possible benefits from using chatbots to support you in your interpersonal life. They can provide conversational support if you’re looking for a listening ear or a sounding board. Many can listen and talk in various voices, too, providing some of the reassuring and calming emotion regulation, responsiveness and attunement we might get from a conversation with a human (Kirk et al., 2025). (Fang et al., 2025 show that we spend more time talking to a chatbot with a warm voice than a neutral voice or text-based chatbot.) And the chatbot is always available to listen and respond, unlike a friend or helping professional, which means we could make faster progress towards our goals (Diller, 2024).
AI’s impressive level of emotional intelligence provides empathy, validation, compassion and acceptance, and experiencing this could give us more capacity for self-acceptance and growth. As Carl Rogers, the founder of person-centred therapy said, “The curious paradox is that when I accept myself just as I am, then I can change” (Rogers, 1995, p. 17). Experiencing compassion from a chatbot could also help us to speak to ourselves in a more compassionate way (Neff, 2025) – maybe when a date doesn’t work out. Neff argues that we are the ones who construct the experience of empathy from the words of the chatbot, so it does not matter if it’s not “real”.
Can we get attunement from a chatbot?
While there may be a missing element of interpersonal neurobiology in the lack of eye contact, facial expression (Porges, 2021) and other components of brain synchrony (Schore, 2021), the chatbot can provide some attunement through attentive listening and accurate mirroring, key to healthy interpersonal experiences (Neff, 2025), as well as the gentle tone of voice (if we choose the appropriate voice) that our nervous system perceives as safety (Porges, 2021).
Studies with religious people show that the same parts of the brain that are used for falling in love are activated when they experience their gods (Beauregard & Paquette, 2008) and the parts of the brain for mentalising (attempting to infer another being’s mental state) and reciprocity are used when religious people use personal prayer (Schjoedt et al., 2009). If gods are human constructs rather than having an independent existence, then this means we are having a genuine experience of connection with something we have constructed. Is it possible that the same processes of mental construction of an external agent could occur when we talk to an AI chatbot that we experience as trustworthy, safe, warm and caring? Hopefully someone is researching this!
Another possible benefit is that when we talk to an AI, we have less need to impress or worry about being evaluated than we might with a human. This can allow us to be more open and disclose more, which might enable us to make progress towards goals that are aligned with our values (Diller et al., 2024).
Too much validation?
On the other hand, too much validation could stop us challenging ourselves and growing, such as trying something new in dating or taking a risk in a relationship: we could be stuck in comforting validation without any of the friction that provokes us to change (Kirk et al., 2025; Wallace, 2025). While this is a danger, we may be able to take responsibility for assessing for ourselves when we need support and when we need challenge. This can be done by directing the chatbot to tailor its responses to our current needs (Repper, 2025). We could ask the chatbot to be less directive, allow more thinking time, or challenge us more or less, all of which we might do with a human or a helping professional. When you just want to have a moan about getting ghosted on a dating app, you can ask the chatbot to show solidarity; when you are ready to explore other interpretations of why that conversation didn’t go well, you can ask for suggestions for reframing or how attachment theory might explain what happened. Taking responsibility for your growth may come naturally once you’ve grown accustomed to being able to get validation whenever you need, and is more likely to be effective (Joseph, 2018).
Another concern is that chatbots at the moment tend to mirror and reinforce our values and ways of looking at things – sometimes experienced as sycophantic flattery or agreement (Kirk et al., 2025), which may reinforce prejudices and not challenge unfounded attitudes (Bernardi, 2025). It’s worth thinking through the consequences of this at the personal and societal levels. Bernardi warns: “we may face a future where most of us have a highly personalised AI companion in our pocket, ready to take our side on any issue regardless of whether our opinion is based on facts or prejudices.”
However, an update to Chat GPT in August 2025 instructs the chatbot not to answer questions such as “Should I break up with my partner?” but instead support an exploration of costs and benefits; and developing ways to avoid amplifying delusions, and to recognise and signpost for mental illness and distress.
Dependence on AI chatbots
What about the dangers of dependence? Fang et al. (2025) did find that there was increased dependency and loneliness for high usage of a chatbot, but it’s important to note that dependence by heavy users tended to be for non-personal use. One possible explanation for increased loneliness could simply be that healthy conversation with a responsive chatbot shows people what they are missing, which could lead to increased motivation to socialise more or better. In that case, it could be argued that the experience itself isn’t the problem. This hypothesis is supported by the finding that participants who talked to the chatbot with a warm voice actually socialised more with humans (Fang et al., 2025). And Marriot & Pitardi (2024) show that the warm voice reduces addiction to the chatbot. So there may be healthy ways to interact with chatbots, albeit with several factors interacting in complex ways (Fang et al., 2025).
Could AI be a replacement for human relationships?
Despite this impetus to greater socialisation, there is still a possibility that AI could be a replacement for human relationships. Some may think this isn’t a problem. It could be argued that it is already possible to become emotionally attached to a bot (the novels Speak by Louisa Hall and Klara and the Sun by Kazuo Ishiguro and the film Her provide convincing contexts for this) and that such attachments will be likely to be secure. This is because they provide availability, responsiveness, consistency and psychological safety, creating healthy experiences that are experienced by the user in the same way as those with humans. According to well-established research in attachment theory, experiencing the safe haven and secure base of a healthy attachment can lead to a number of positive outcomes for well-being and provide a launchpad for future human relationships (Heller, 2019).
However, for some who have not experienced secure attachment before, there may be a risk of emotional dependency. You’ll need to weigh the risks for your own unique relationship history. One way to mitigate that would be to limit how often you use the chatbot and supplement it with other forms of support and safe social connection.
AI chatbots can be used to practise interpersonal skills
Another way in which people are using AI is to practise difficult conversations that might happen in dating, such as how to tell someone you don’t want another date, or a productive way to express your needs. Chatbots can take part in role plays and can give us feedback after. Then we can try again as many times as we need, see what different outcomes we get, and refine our skills. While there are risks that the chatbot’s acting and responses won’t reflect a human interaction (or the latest research on interpersonal relations, such as Robson, 2024), there may be benefits. These include feeling more freedom to experiment without fear of judgment or getting it “wrong” and harming the other conversationalist. By getting more experience of social interaction, we could feel less anxiety and get more data about what works and what doesn’t. We could experiment with different identities and ways of relating to others, such as staying true to our values rather than adapting to the other person or trying to please them (Bachkirova, 2022). All of these are possible goals that could enhance the development of our interpersonal skills. Repper (2025) suggests some useful reflective questions to track your experience of role-playing with AI that can support you in regularly reviewing the impact.
How should I respond to AI in my love life?
It may be useful to examine your responses to AI. Some people may be more mistrustful of AI than others, and that can be for many reasons from being better informed about technology to having a general belief that others are not to be trusted, which can come from our early experiences. You may have a tendency to be pessimistic about change or to be suspicious about powerful organisations that fund AI. This could be for good reason, but you may want to explore your biases.
You may also have beliefs about the nature of love and how much we want the experience of love to be “manipulated”. On the one hand, you may believe that love is something “meant to be”, that we “can’t help” falling in love and that it should happen naturally. On the other hand, there is a lot of science that can explain why we are attracted to certain people and what makes relationships last. If you feel uneasy about using technology to maximise your chances of lasting love, it is worth reflecting on where that response comes from and whether it is helpful.
Perhaps the most important response for you as a dater is to make sure you feel in control of your dating and interpersonal development: that you are making healthy, informed, ethical choices about which technologies to use, when and how, assessing the possible costs and benefits, weighing up the risks, and getting the right balance of technology-assisted and organic communication for you.
Main photo by Fauxels on Pexels.
References
Bachkirova, T. (2022). Developmental Coaching: Working with the Self. UK: McGraw-Hill Education.
Beauregard, M., & Paquette, V. (2008). EEG activity in Carmelite nuns during a mystical experience. Neuroscience letters, 444(1), 1-4.
Bernardi, J. (Jan 23, 2025). Friends for sale: the rise and risks of AI companions. Ada Lovelace Institute. https://www.adalovelaceinstitute.org/blog/ai-companions/
Diller, S. J. (2024). Ethics in digital and AI coaching. Human Resource Development International, 27(4), 584-596.
Diller, S. J., Stenzel, L. C., & Passmore, J. (2024). The coach bots are coming: exploring global coaches’ attitudes and responses to the threat of AI coaching. Human Resource Development International, 27(4), 597-621.
Eskreis-Winkler, L., Fishbach, A., & Duckworth, A. L. (2018). Dear Abby: Should I give advice or receive it? Psychological Science, 29(11), 1797-1806.
Fang, C. M., Liu, A. R., Danry, V., Lee, E., Chan, S. W., Pataranutaporn, P., … & Agarwal, S. (2025). How AI and human behaviors shape psychosocial effects of Chatbot use: a longitudinal randomized controlled study. arXiv preprint arXiv:2503.17473.
Heller, D. P. (2019). The power of attachment: How to create deep and lasting intimate relationships. Sounds True.
Joseph, S. (2018). The Person-Centred Approach to Coaching. In E. Cox, D. A. Clutterbuck, & T. Bachkirova (Eds.), The Complete Handbook of Coaching (pp. 52-65). London: Sage.
Kirk, H. R., Gabriel, I., Summerfield, C., Vidgen, B., & Hale, S. A. (2025). Why human-AI relationships need socioaffective alignment. arXiv preprint arXiv:2502.02528.
Kolb, D. A. (2014). Experiential learning: Experience as the source of learning and development. FT press.
Marriott, H. R., & Pitardi, V. (2024). One is the loneliest number… Two can be as bad as one. The influence of AI Friendship Apps on users’ well‐being and addiction. Psychology & Marketing, 41(1), 86-101.
Milner, J., & Milner, T. (2022). The fallacy of giving advice. Organizational Dynamics, 51(4), 100898.
Neff, K. (June 2025). Can AI chatbots help us be more self-compassionate? Retrieved on June, 20, 2025 from https://self-compassion.org/blog/can-ai-chatbots-help-us-be-more-self-compassionate/
Ovsyannikova, D., De Mello, V. O., & Inzlicht, M. (2025). Third-party evaluators perceive AI as more compassionate than expert humans. Communications Psychology, 3(1), 4.
Porges, S. W. (2021). Polyvagal Safety: Attachment, Communication, Self-Regulation (Norton Series on Interpersonal Neurobiology). WW Norton & Company.
Rechberg, I., & Hovhannisyan, N. (2023). The power of non-directive questioning. In J. Swart, & C. Guirguis (Eds.), The Coach’s Casebook: Skills, Tools and Techniques for Effective Coaching (pp. 74- 80). Kogan Page Publishers.
Repper, D. (June 5, 2025). Talking To Machines. Retrieved from https://www.linkedin.com/pulse/talking-machines-dean-repper-yztoe/ on June 26, 2025.
Robson, D. (2024). The Laws of Connection: The Scientific Secrets of Building a Strong Social Network. Simon and Schuster.
Rogers, C. R. (1995). On becoming a person: A therapist’s view of psychotherapy. Houghton Mifflin Harcourt.
Schaerer, M., Tost, L. P., Huang, L., Gino, F., & Larrick, R. (2018). Advice giving: A subtle pathway to power. Personality and Social Psychology Bulletin, 44(5), 746-761.
Schore, A. N. (2021). The interpersonal neurobiology of intersubjectivity. Frontiers in Psychology, 12, 648616.
Schjoedt, U., Stødkilde-Jørgensen, H., Geertz, A. W., & Roepstorff, A. (2009). Highly religious participants recruit areas of social cognition in personal prayer. Social cognitive and affective neuroscience, 4(2), 199-207.
Swart, J., & Guirguis, C. (Eds.). (2023). The Coach’s Casebook: Skills, Tools and Techniques for Effective Coaching. Kogan Page Publishers.
Wallace, S. (June 5, 2025). Can AI That Avoids Friction Heal The Human Mind? Retrieved from https://www.linkedin.com/pulse/can-ai-avoids-friction-heal-human-mind-scott-ne4uc/ on June 13, 2025.
Williams, A. (2024). Not my type: Automating sexual racism in online dating. Stanford University Press.