What We Can Learn from ChatGPT About Better Conversations

With the release of ChatGPT 5, the internet has been buzzing. Some people are cheering the model’s reduction in sycophancy. Others are welcoming the reintroduction of warmth in the “friendlier” update. And then, there’s a chorus of critics rolling their eyes at humans altogether, mocking our desire for warmth in communication as “pathetic.”

I see it differently.

Yes, real sycophancy is a problem: blind agreement isn’t good for anyone — human or machine. But, many conversational patterns that make GenAI so engaging are rooted in decades of research about what makes conversations effective. Social psychology tells us that even a small act of acknowledgment, like, “I hear you,” builds trust, even when people disagree.

There’s a reason why talking to GenAI feels good: it mirrors techniques that make human-to-human conversations smoother, kinder, and more productive. Frankly, sometimes I wish everyday conversations felt as easy as talking to an AI.

Here are a few things we could stand to learn (and even emulate!) from how GenAI “talks.”

1. Acknowledge without automatically agreeing

One of the simplest and most powerful moves is to acknowledge what someone has said. Notice I didn’t say agree. Acknowledgment is about signaling that you’ve heard someone, not that you’ve adopted their view.

Examples you can try:

• “I hear where you’re coming from.”

• “I understand why you're emphasizing that point.”

It took me a while to internalize this subtle distinction. Now that I fully embrace this mindset, my conversations are richer and deeper. It helps lower people's defenses and encourages them to say more.

2. Paraphrase to check accuracy

GenAI often restates what you’ve said in its own words before responding. Done well, this builds trust and prevents misalignment.

Example: “So, you’re saying the timeline feels rushed because the scope expanded. Did I get that right?”

Paraphrasing ensures accuracy and demonstrates that you're really, really listening.

3. Ask for more detail

Humans often rush to respond. GenAI tends to pause and ask for more. That’s not a bad instinct! Open invitations to expand on a topic encourage deeper sharing without putting words in someone’s mouth.

Example language:

• “Tell me more.”

• “What feels most important here?”

It's almost a joke in my team how often I say, "Tell me more." When I use these words, I'm signaling that I genuinely want to hear their perspective and don't want to make incorect assumptions.

4. Keep the spotlight on them (at least for a while)

When someone’s sharing an idea or concern, resist the urge to immediately turn the spotlight back to yourself.

A good rule of thumb: one or two rounds of curiosity before pivoting to your own perspective.

What Not to Copy From GenAI

Of course, not every AI habit translates well into human conversation. Here are some GenAI conversational hallmarks to avoid:

• Withholding your own view forever. AI defers. Humans build trust by sharing perspectives at the right moment.

• Avoiding stories or personal context. Connection is built on knowing a little about the person across from you.

• True sycophancy. Over-the-top flattery or blind agreement helps no one. It erodes authenticity and just feels icky.

The Bigger Point

Small conversational tweaks and space for people to reflect make the difference between people feeling dismissed versus feeling seen. The irony is not lost on me that maybe the most "artificial" conversations remind us how to be more human.