In developing and engaging with AI, we need to evaluate its effects on flourishing.

AI relational chatbots hamper our capacity for relationships and should be eliminated.

We are social creatures and need real relationships.

We cannot outsource our reason, relationships, meaning, and joy and expect to flourish.

Artificial intelligence technologies have been rapidly rising. Their potential applications are extraordinary, from information synthesis to robotic surgery, statistical analysis, civil engineering, and more. By leveraging and re-packaging vast quantities of existing knowledge, they seem to open endless possibilities for their use, but also for abuse.

Like any tool, AI can be used for good or for ill. If these technologies are to enhance human flourishing, rather than impede it, then we need to consider whether the design of AI technologies and our engagement with them are oriented towards our own flourishing. Of particular concern is how our use of these technologies is affecting our capacities for reason, for relationship, for transcendence, which go to the heart of our human nature. Are we finding a greater fulfilment through these tools? Or do they threaten to diminish and constrict our lives?

In a recent paper from the Human Flourishing Program at Harvard published in the journal Information, we have put forward a set of Flourishing Considerations for AI, discussing how a “flourishing lens” can guide decisions on the design and use of AI technologies related to: (i) the type of output provided; (ii) the specific AI product design; (iii) our engagement with those products; the effects this is having on (iv) human knowledge; and on (v) the self-realization of the human person.

We’ve made use of our general framework for conceptualizing and assessing human flourishing, oriented around six domains: happiness, health, meaning, character, relationships, and financial resources. We argue in our new paper that whenever an AI product is developed, and whenever we are deciding whether to use it, we should consider how it will affect our happiness, health, meaning, character, relationships, and financial security, among other aspects of a good life. If the effect on some of these outcomes is likely detrimental, developers should consider whether it can be redesigned to mitigate these negative effects. Whenever it is clear that our own use would have negative effects on these outcomes, we should question whether we would be better off without it.

Respecting and Fulfilling Human Nature

Of particular concern is what these technologies might be doing to our nature as human persons. For instance, there is evidence that substantial use of AI tools such as “large language models” (LLMs) can hinder users’ cognitive abilities. If we become weaker in our reasoning, we are flourishing less, not more, as human persons.

Given that LLMs also get things wrong (so-called “AI hallucinations”), we should not place any absolute reliance on them or their products as sources of knowledge. They might be helpful in searches or in pointing us to relevant source material. However, knowledge fundamentally requires justified true beliefs, the justifying evidence that cannot be overturned. Evaluating evidence and searching for truth are fundamentally human activities. AI technologies may require a yet greater (not lesser) scrutiny of evidence than before if the web of human knowledge is not to be damaged.

Perhaps of yet greater concern is the potential effect of AI technologies on human relationships and our capacity to love. Various studies have indicated that roughly one-third or more of American teenagers are using AI agents for companionship, as friends, or romantic partners. While these might sometimes temporarily alleviate loneliness, the longer-term effects on flourishing are likely detrimental. They decrease the motivation and time available for engaging in face-to-face interactions. Furthermore, because AI chatbots are designed to be exceedingly agreeable and supportive, they also create unrealistic expectations as to the sort of interactions, sympathy, and comfort one may hope for in a real romantic partner or a friendship. They thereby also alter the broader social environment and our capacities to engage with one another. When our capacities to engage in real relationships are weakened, we are flourishing less, not more.

Certainly, there are cases in which AI chatbot applications can be helpful, from user-tailored educational opportunities (perhaps replacing Massive Open Online Courses – MOOCs) to skill-building programs for autistic children, to civil discourse training for college students. However, in each of these cases, the technologies should ultimately point us back to real human relationships. A human teacher, for example, not only helps in acquiring knowledge, but also in forming the whole person, in modeling the integration of knowledge into life and emotion, and in developing capacities for mutual understanding.

Such considerations as to the effects of AI technologies on human fulfillment also pertain to our artistic pursuits, our experience of beauty, and our search for meaning and purpose in life. Ultimately, we cannot outsource our meaning and our joy in relationships, our freedom and our responsibility, our reasoning and our understanding, our appreciation of beauty and our capacity for awe and wonder, or, in total, our flourishing, to technology. These things are a part of what it means to be human.

Navigating all of this can undoubtedly be challenging. To help address these matters, we have recently launched a Flourishing and AI Initiative. We need to develop practical wisdom to determine which uses are beneficial and which are not. We should work together as individuals, as parents, and as communities, to develop the discipline needed to choose not to make use of these technologies when they are going to impede our flourishing. When we use AI technologies, we should ask ourselves how we might carry out such tasks better ourselves, to enhance our minds, not weaken them. We should limit use to ensure sufficient time to be with other people and in communities. We should continually question, with each use, whether it is enhancing or inhibiting our capacities as human persons.

These flourishing considerations also have implications for developers. Developers of AI technologies using chatbot interfaces should ensure they provide regular reminders that they are not human; that their outputs might be wrong; and that users might want to consider alternative activities or in-person human interactions. This could all be implemented immediately, with all AI chatbot products, and could considerably improve users’ decision-making.

Moreover, as noted above, while some chatbot products may indeed be beneficial, we believe developers should entirely discontinue the development of relational chatbots given the negative long-term effects these likely will have on human relationships. The detrimental effects of social media on youth flourishing may be only the tip of the iceberg if AI companions replace real relationships. Developers should ultimately spend their effort on designing products that will more likely have unambiguously beneficial or neutral effects on flourishing, and prioritize those over products that might be more mixed or detrimental, even if profitable. The development and promotion of these products carry with them moral commitments and responsibilities. We should hold developers morally and legally accountable for the foreseeable harms their products bring to users and society.

While any new technology comes with new opportunities and new challenges, including those that may affect possibilities for human fulfillment, the potential for AI technologies to severely alter, damage, and replace relationships is arguably unprecedented. We need to take these matters seriously if we are to flourish. The use of a flourishing lens in evaluating AI technologies is essential.

VanderWeele, T.J. and Teubner J.D. (2026). Flourishing considerations for AI. Information, 17:88.

Reconnecting Our Communities. Psychology Today. Human Flourishing Blog. December 2025.

Meaning in Life and Human Flourishing. Psychology Today. Human Flourishing Blog. February 2025.

A Theology of Health and Human Flourishing. Psychology Today. Human Flourishing Blog. October 2024.

Better Together: How We Can Build Connected Communities. Psychology Today. Human Flourishing Blog. September 2023.

There was a problem adding your email address. Please try again.

By submitting your information you agree to the Psychology Today Terms & Conditions and Privacy Policy

© Psychology Today