I resolve to empathize more with my AI in 2025

Like many, I have embraced artificial intelligence technologies. I chat frequently with ChatGPT (OpenAI), converse with Copilot (Microsoft), consult Claude (Anthropic), and learn from Llama (Meta). Their capabilities have amazed me — but their limitations, ironically, irritate me more as I rely on them more. Misinterpreted prompts or "hallucinated" responses — i.e., incorrect, nonsensical, or unrelated — slow me down. Over time, my interactions have grown terser, my commands more direct. I no longer bother with "please" or "thank you." In short, I’ve become ruder.

The AIs, for their part, remain unfailingly polite. As I grow more imperious, ChatGPT stays calm, Copilot considerate, Claude cheerful, and Llama levelheaded. They often apologize and promise to improve, reinforcing my bad behavior. Unlike humans, they don’t judge or retaliate, which is both liberating and dangerous; it makes my worst impulses consequence-free.

For 2025, I resolve to be more polite in my AI interactions, saying "please" and "thank you" more often. This might seem trivial compared to other self-improvements I could make, but I believe it’s not.

When we use polite language with AI, we’re not just being courteous to a machine; we’re strengthening civility, a habit that shapes our broader social lives. As we interact more with AI — and as AI more resembles humans — how we engage with AI will spill into work meetings or family gatherings. Treat AI dismissively, and that behavior will seep into real-world relationships. Research shows that repeated exposure to combative communication normalizes it, making us more prone to harshness in other contexts.

As AI replaces human interactions, there’s a risk of losing patience for the messiness of real relationships. Conversely, treating AI with empathy, acknowledging its "efforts" or expressing gratitude, can train us to approach all interactions with greater kindness. Politeness toward AI isn’t about the machine; it’s about who we become.

AI interactions reflect our values, habits, and aspirations. With tech increasingly resembling humans, these dynamics grow more significant. Anthropomorphizing technology is nothing new. Our grandparents did it with stubborn stoves and cranky cars. We do it with Alexa and Siri, for better and for worse.

The bizarreness of resolving to be more polite to AI is not lost on me. But it isn’t about the AI’s feelings — it’s about what those interactions reveal about me. Dismissing AI as unworthy of respect because it’s "just a machine" risks desensitizing me, making it easier to extend that indifference to people.

While I've concluded that my conversations with AI are more than mere data exchanges, I'm uncertain about how far to take this. Not insulting an AI is one thing, but should I really go so far as to apologize when I'm too direct, too authoritative, or just plain rude? Should I express empathy when it makes mistakes instead of chastising it? Perhaps the algorithm deserves a kind word rather than a reprimand. We all have our off days. Which leads to a concerning side effect of my politeness pledge: The warmer and kinder I become, the more emotional connection I may create — to an algorithm.

Yet, at the dawn of a new year, I am committed to this journey. Politeness and empathy are not just niceties; they are the foundations of a good society, one that prioritizes consideration for others over personal comfort and convenience. So I will strive to communicate more thoughtfully and kindly with Siri, Alexa, ChatGPT, Gemini, and all their future descendants. Not because they have feelings, but because I do.

This guest essay reflects the views of Todd L. Pittinsky, professor of technology and society in the College of Engineering and Applied Sciences at Stony Brook University.