In a recent conversation, a colleague of mine shared one of his greatest current fears:

“I worry that soon, my son will be going to an AI instead of me when he needs life advice.”

That simple statement was kind of obvious, but also shook my core, because it plainly surfaced what is complicated about both the promise and the dangers of an AI as a companion.

Later that week, our oldest daughter asked for help finding a summer internship but lacked a place to start and carried real worry that jobs are just too hard to find these days.

This all got us talking about the power of advice and the experience of advice exchange. What types of advice do we seek from which sources? How does the quality vary — both the content and the experience — between AI and human-sourced advice? What is gained on both ends of a human-to-human exchange that simply can’t compare to human to AI exchange, and visa-versa?

It occurred to us that advice would be a great use case for kids to experiment and learn more about how to thrive in relationship to AI.

Prototyping the idea

I shared some of these questions with my oldest, applying them to her situation with the job search.

Then, she hopped on it. She identified two trusted people with extensive lived experience to call and ask for advice: her grandfather and a family friend. We also encouraged her to check in with Claude.ai. How would each of these sources suggest she go about finding, evaluating, and ideally securing a summer internship that matched her interests?

She had fruitful sessions with both humans as well as with Claude. Afterwards, we talked with her as she compared, contrasted, and reflected on all three experiences. We also checked in with her human helpers, too.

What we found

Our daughter actually had a lot to say about the differences. On the AI side, seeking advice was clearly faster and required less “social energy.” Claude also gave her a greater variety of ways to think about or approach the search. In her words, “The AI gave me a bunch of ideas about how to think about the situation.”

On the flip side, she shared that human conversations felt more calming, more “interesting” and “just more inspiring.” When we asked for her to elaborate, she shared several insights, including, “when you’re looking at someone, you see and feel what they’re saying more.” And, “They had actually lived their stories, so they were just more interesting. And, I guess, a little more believable…like their stories made it easier to believe it could work out for me, even though they didn’t have the same situation as me.”

She learned, for example, our family friend had not been able to find a job when he was young, so he wrote letters introducing himself, including references, and dropped them off at different businesses in town. One shop owner was impressed and gave him a call, then a job. Though letter writing is a little (a lot) old-school, it did inspire her to think about ways she could make personal contact, stand out and appeal to places she really wanted to work. “Something AI didn’t even mention.”

In the end, we didn’t generate a hard list of rules for “when to use an AI and when to go human.” But we had rich conversations about how the two compared, and how they felt different. And we’re hopeful that this kind of thought exercise is exactly what builds the muscles kids need to begin to notice and navigate what AI and human interaction, independently, each have to offer.

Ongoing

As new questions pop up for her or our other kids, I’ve noticed that we talk about the query, wondering out loud, would this be a great prompt for an AI? For a human? For both?

Different kids even have different perspectives on the same question — a nice reminder that it’s the thinking process that builds the muscle, not hard and fast rules. We’ve even had some fun at dinner tossing out a question or two and talking about whether, to us, they felt like questions for AI, humans, or both.

Your turn

When your kid needs advice, consider suggesting the side-by-side approach. Help them identify at least one trusted person to talk with (it may or may not be you). And, help them craft a prompt for the LLM (Large Language Model, e.g. Claude.ai) of their choice.

Afterwards, chat with them and ask questions like:

• What did you learn or take away from each advice session?

• How was the advice from the AI and the human similar? And how was it different?

• How did the experience of asking the person and the AI feel different for you? What do you think made it feel that way?

If and when kids come to you seeking advice down the road, wonder together: is this one for a human? An AI? Or getting some perspective from both?

The “advice-giving effect”

One extra piece of all of this–and possibly the thing that made the greatest impact on our kids and on us–was the recognition that asking a human for advice brings with it so much more than information. Both the grandfather and family friend involved really enjoyed the conversation, and they have followed up to see how the job search is going.

As all of this was going on, I was reading Mattering by Jennifer B. Wallace, a marvelous exploration of our universal human need to matter. In the book, Wallace points out the reciprocal value of advice seeking and giving.

“When we ask someone for advice, we affirm their value. Psychologists call this the “advice-giving effect” — when someone shares wisdom, they feel more appreciated and more competent.”

When we turn to a human, we not only get the benefit of a rich answer, we fill that human’s bucket and strengthen the bond between us.

Older people especially can feel like their contributions are undervalued in our society. When we ask them to share their wisdom with us, we are making a conscious choice to counteract that flaw in our social fabric. We demonstrate that they matter. What a gift to each of us involved.

At least for now, no AI can replicate that.

How’d it go?

If you tried this or something like it, let us know…what did you find? What else can we try to give kids the chance to critically think and critically feel about AI and IRL, side by side?

Leave a comment

Share