In recent years, Artificial Intelligence (AI) has become an integral part of the daily lives of millions of people around the world. Studies, surveys, and experiments conducted across different countries suggest that its impact is far from straightforward.

A recent ongoing study by researchers from the MIT Media Lab and OpenAI followed 981 participants over a four-week period and analyzed more than 300,000 interactions with chatbots.

The findings challenge several common assumptions about AI:

Moderate AI use was associated with a slight reduction in feelings of loneliness.

However, increased daily use was consistently associated with:

• • • Higher levels of loneliness,
    • Reduced social interactions with other people,
    • Stronger emotional dependence on AI,
    • A greater risk of problematic use.

One of the most surprising findings was that:

Voice-based interactions with AI were not found to be psychologically riskier than text-based interactions. In fact, during the early stages of use, voice conversations were associated with lower emotional dependence and less problematic use than text conversations. However, these benefits diminished as daily usage time increased.

Another unexpected insight:

Heavy users who engaged in practical conversations with AI (such as seeking advice, explanations, or problem-solving assistance) sometimes demonstrated stronger emotional dependence than users who discussed personal topics.

The central conclusion is not that AI is inherently good or bad for people.

The real question is how we choose to use it.

When used as a tool, AI can enhance productivity, learning, decision-making, and critical thinking.

When used as a substitute for human relationships, independent decision-making, or personal self-efficacy, it may gradually create new forms of dependency.

As AI becomes a daily companion for millions of people, AI literacy will need to include not only technical skills, but also an understanding of the psychological and social effects of human–machine interaction.

Research increasingly suggests that the primary challenge of the future is not artificial intelligence itself.

The real challenge is managing our relationship with artificial intelligence.

Source:

Fang, C.M., Liu, A.R., Danry, V., Lee, E., Chan, S.W.T., Pataranutaporn, P., Maes, P., Phang, J., Lampe, M., Ahmad, L., & Agarwal, S. (2025). How AI and Human Behaviors Shape Psychosocial Effects of Chatbot Use: A Longitudinal Randomized Controlled Study. MIT Media Lab & OpenAI.

A randomized controlled study involving 981 participants, more than 300,000 messages, and a four-week research period, examining loneliness, socialization, emotional dependence on AI, and problematic AI use.