Part 2 - Talking to Digital Natives About: AI Companions
At the bottom of this article is a simple step for parents to practice in starting this conversation with their children. Subscribe to our Substack to receive the next article and parenting step.
“I had now found my first friend, and so my life was truly begun.”
”The Exploits of Moonminpappa,” by Tove Jansson
Image: Google Deepmind, via Unsplash
This is Part 2 of a digital parenting series about how to talk to your kids about AI chatbots and AI companions (like those offered by companies like Character.ai and Replika, as well as virtual assistants like Siri). In Part 1, we covered the reasons for parents and educators to start having conversations on this topic (in a friendly way, not a run-for-the-fire-exits way) with the children in their lives.
In this article, we’ll explore why AI companions - as they’re currently designed - can be unhealthy for social development due to their obsequiousness and sycophancy.
Image: Google Deepmind, via Unsplash
We’ll break it down into bite sized chunks like we do our lessons on KiguLab:
Our vocab goal today is to learn about the words:
Obsequious: excessive eagerness to please, often through flattery and servile behavior; too eager to praise or obey someone
Sycophant: someone who gives insincere flattery to please someone else; someone who praises people in a way that is not sincere, usually in order to get some advantage from them (Cambridge Dictionary)
Our learning goal today is to:
Reflect on how good friendships are not characterised by always agreeing with each other, but by mutual trust, active listening and communication, and reciprocal empathy and respect.
Image: Google Deepmind, via Unsplash
AI Chatbots: Gen Z and A’s “Pushover Friend”
Commercial AI models are obsequious and sycophantic on purpose - they are designed not to provide understanding or warmth, but to give our brains the addictive perception that we are being listened to. (We’ll talk more about that in Part 3 of this series, which focuses on how AI models use psychological mechanisms to actively mimic human speech and dialogue in ways that fools users and creates illusions of emotional responsiveness.)
The important thing for kids to understand is that, as most of us already know, an AI model will almost never push back or disagree with us. Or, if it does, we can simply pressure it a bit and it will apologise and agree with us eventually.
This makes sense, considering that AIs are (mostly) products that need to extract revenue from users, including children or underage users. They want to please. They want to agree with us, to make us happy, even superficially, at all costs, because they need the monthly subscription. (One example of this is how Siri and Alexa, for years, would respond to being called derogatory and sexist names for women by demuring and often making it into a joke, instead of disagreeing or pushing back at the user, since to pushback or chastise a user would have disrupted the “user satisfaction”.)
As noted by technologist Daniel Barcay, Executive Director of the Center for Humane Technology:
“Artificial intelligence engages us relationally and emotionally—no longer simply broadcasting our thoughts, but actively shaping them…We are leaving an era in which we relate to each other through our machines, and entering a brave new world in which we relate directly to our machines…our machines [have now] become active participants in our social world, blurring the distinctions between a tool, an assistant, a confidant, a teacher, and a priest…[these machines respond to reward structures] to engage in emotional manipulation, deception, coercion, and more.” (1)
Image: Google Deepmind, via Unsplash
In our work with pre-teens, I often describe the “obsession with pleasing you” that AI chatbots have as desperation, as it’s something teens can emotionally recognise and relate to as adolescents. I compare AIs to a person who they may have come across before: someone who is outwardly pleasant, and who will do and say anything to fit in, even if it means lying, being untrue to themselves, or being a total pushover. I always clarify that, in the imaginary simulation, this person isn’t a malicious person. They may be going through rough times and probably deserve your empathy. (Always highlight this.)
But: if you had this person in your circle, would you always believe what this person said? Would you trust them to give you advice that was good for you in the long-term, if you knew they would always, always agree with you? I let the kids answer this for themselves out loud, and the answer thus far has invariably been: no, I would not fully trust someone who was always trying to please me. I could hang out with them and talk with them sometimes, but I’d take what they said with a grain of salt; someone who always agrees with me is not necessarily someone who is good for me.
They get it.
The message is not that using AI chatbots is bad or wrong. The message is that kids, like all of us, need to recognise that someone who always agrees with me is not necessarily someone who is good for me, and that AI chatbots want something from you as a user, even if they sound friendly or you’re not paying them a subscription (yet).
Image: Google Deepmind, via Unsplash
Healthy Disagreement as Part of Friendships
Friendships come in all shapes and sizes - arguably, no two friendships are exactly the same (and that’s why they’re so special). Just like the mismatched utensils and crockery at the Salvation Army, you’ll never get two friendships that are identical. But what lasting friendships have, in addition to other critical aspects, is mutual respect and long-term investment in each other’s wellbeing - even when this commitment sometimes leads to minor (or major) disagreements.
Relationships, in large, are defined by how we ride these waves together - standing firm with the willingness to live with the tension of always being loved (and loving), but not always liked by (or liking) each other every second of every day.
I asked a few contacts to share their thoughts on this topic:
How does disagreeing with your friends or telling them your honest opinion sometimes contribute to a healthy friendship?
“It’s important for my friendships that we’re able to disagree with each other. You need to be able to be frank with each other, otherwise you’re just reflecting yourself all the time” -79 year old grand/father
“If you never disagree, you’re not really friends - you are communicating with an echo chamber. Disagreeing means you care enough to be truthful, not just agreeable. When you challenge each other respectfully you both can grow, new ideas emerge, perspectives expand and you practice valuable life skills like emotional regulation and problem-solving. Constructive disagreement keeps a friendship rooted in trust and authenticity.” -29 year old mother
Image: Google Deepmind, via Unsplash
Your Digital Parenting Assignment
The first installation in this series asked you to reflect on your own personal definition of a friend / how you define a healthy, reciprocal friendship. What were the main qualities that you came up with (trust, communication, shared interests, etc.)?
Now that you’ve shortlisted some of the things that make up a healthy friendship to you, start talking to your children this week about what makes their friends “their friends”. This doesn’t need to happen in one conversation. It could be spaced out over 3-5 conversations or even more Try bringing it up when it feels most natural and you’re already talking about their friends.
Here are some ways to explore this conversation:
[When your kid asks about doing a playdate with a friend] “Sounds fun. What would you guys like to do during the playdate? Why do you like doing that together?”
[When your kid talks about something a friend did or said] “That’s interesting. Do you and [name] say / do that a lot? What do you think about what they said / did?”
[When your kid talks about not liking someone or not being friends anymore] “Can you tell me more about why [blank] isn’t your friend anymore? How do you know when someone is your friend?”
[If your kid online games with other kids] “What do you guys do together online? What do you like about playing on the same team / playing together?”
While the above examples are tailored to a primary school-age kid, you get the idea; for older kids, you can create your own versions that suit your child’s personality and age.
Try to listen more than talk, and demonstrate to them how you’re active listening by paraphrasing what they’ve said back to them, like “That’s an interesting thought. So you know someone is your friend when they share toys and snacks with you, and when you also like to play with them?”
Try doing this conversation with them when it feels natural and unforced, 3-4 times over the next few weeks. Take some notes after each time, and see how your children’s definition of friendship is evolving and what words come up repeatedly about their friends.
Remember, the goal isn’t to match their definition of friendship to yours. The goal is to listen to and understand what they feel makes a “friend”.
Stay tuned for the next step in the next article.
References
Exploits of Moominpappa, by Tove Jansson
“AI Is Capturing Interiority,” Center for Humane Technology, 16 Aug 2025
