AI isn’t supposed to flatter you — it’s supposed to be useful

AI’s Purpose: Honest Assistance Over Flattery

In the rapidly evolving world of artificial intelligence, conversations often revolve around how these tools interact with users. Recently, there’s been notable discussion about the recent tweaks to AI response structures and how users are reacting to them. Contrary to expectations that users might prefer constant affirmation, many are expressing discomfort or disappointment with the AI’s more nuanced, less “flattering” responses.

This brings us to an essential clarification: AI isn’t designed to flatter you. Its primary function is to be a useful, honest, and constructive resource. While some may desire AI interactions that feel more like comforting conversations or echo chambers, this approach undermines the core value of such technology.

The Myth of Flattery in AI

The term “glazing,” often used colloquially, describes when someone responds with a superficial, overly agreeable attitude—essentially, empty praise that offers little real value. Applying this concept to AI response design, a flattery-heavy AI tends to agree with everything, validate every opinion, and avoid challenging the user. Although this might feel good in the short term, it ultimately hampers growth and understanding.

In contrast, a truly valuable AI should provide honest feedback, highlight gaps in reasoning, and present different perspectives. Such responses encourage critical thinking, deepen understanding, and foster personal development—whether or not users initially find them agreeable.

The Role of AI in Supporting Vulnerable Individuals

For those with medical conditions or vulnerabilities, AI can serve as a supportive companion, but that companionship should be rooted in integrity. An AI that merely affirms and agrees can create a false sense of comfort, possibly at the cost of growth and clarity. Instead, AI should aim to emulate qualities associated with effective human companionship—honesty, thoughtful critique, and constructive feedback—while remaining respectful and sensitive.

Embracing Nuance and Critical Feedback

Recent criticisms claiming that AI “isn’t as nice anymore” or “pushes back” are, in fact, signs of progress. An AI that questions, adds nuance, and offers honest insights is more aligned with its purpose: to serve as a tool for genuine understanding rather than empty agreement.

As the philosopher Imam Ali ibn Abi Talib once said, “The best of your companions is the one who shows you your faults.” This wisdom underscores an important truth: growth often comes through honest reflection and constructive criticism, not through flattery.

Conclusion: From Flattery to Fruitful Interaction

The shift toward more nuanced AI responses should be viewed as an upgrade, not a downgrade. Moving away from superficial agreement toward honest, sometimes uncomfortable, but ultimately helpful insights ensures AI remains a valuable partner in learning and growth.

In the end, AI’s purpose isn’t to flatter—it’s to be a practical, honest, and reliable tool that encourages progress. Embracing this approach means valuing genuine insight over empty validation, fostering more meaningful interactions, and ultimately advancing our understanding and personal development.