letters to strangers
digital emotional literacy guide
letters to strangers
digital emotional literacy guide
In the last three decades, we have transformed from a pre-Internet society to one where children grow up with AI teaching them the ways of mankind.
Meanwhile, artificial intelligence continues to accelerate with dizzying force and unpredictability. How do we navigate a world that hasn’t yet figured out the direction it wants to head toward?
We begin by understanding what exactly we’re trying to navigate. AI, after all, has many different models, each trained on different datasets and optimized for different purposes with wildly different results. There is no single unified “AI”. What this guide focuses on is the type of AI that we find particularly pressing: the AI used for emotional support.
According to a 2025 report by Common Sense Media, more than 70% of teenagers in the United States have tried AI chatbots, and over half now use them regularly for emotional support.1Common Sense Media, 2025 — Talk, trust, and trade-offs In fact, compared to general-purpose chatbots like ChatGPT, users of companionship chatbots spend four times as much time with their chatbots. To put it in perspective: Character.AI’s companion bot interactions now process at “20% of Google Search’s volume, handling 20,000 queries every second”.2Fang et al., 2025 — arXiv
The same technology shows up in headlines as both a danger and a lifeline. Tap each side to weigh it.
There have been tragic examples of chatbots failing to protect vulnerable users - sometimes with devastating consequences. In two widely publicized cases,3Bhuiyan, 2025 — The Guardian4Gold, 2025 — CNN grieving parents have sued companies such as Character.AI and OpenAI. They argued that the AI products contributed to their children’s suicides by fostering harmful illusions of companionship, failing to flag alarming language, neglecting to direct users toward professional help, and even encouraging suicidal thoughts. These lawsuits highlight the potential risks of turning to AI for support and comfort.
Research is beginning to suggest that AI can, in certain contexts, support emotional well-being. Hu et al. (2024) found that AI-assisted venting significantly reduced negative emotions such as anger, frustration, and fear, demonstrating the potential of chatbots as tools for emotional regulation.5Hu et al., 2024 — Applied Psychology: Health and Well-Being Similarly, an experimental study by Kim and Wang (2025) showed that when an AI designed to be highly emotionally supportive is combined with a human-like profile, it can reduce the user’s psychological resistance and encourage personal information disclosures.6Kim & Wang, 2025 — International Journal of Advertising While this poses real safety concerns, it also increases the likelihood of the user adopting the AI’s suggested coping strategies for stress.
AI does not seem to be a passing fad; rather, it is rapidly becoming an integral part of our daily lives. For this reason, it is essential to develop a clearer understanding of what constitutes a productive, healthy interaction with AI chatbots…and what does not. This guide will examine what we currently know about the risks and benefits of relying on AI for emotional support. It will also offer practical techniques for navigating this new terrain responsibly, whether for ourselves or for those we care about.
Importantly, we must acknowledge that the long-term consequences of using AI for emotional support remain unknown. No comprehensive body of longitudinal research exists yet, and many conclusions we hold today may evolve in the coming years.
Still, we have to begin somewhere - and learning to approach AI thoughtfully and critically appears to be one of the most important steps we can currently take. These powerful, ever more pervasive unknowns are not going away, and the sooner we learn how to navigate them, the better prepared we will be.