- You are here:
- Home
- Maitri Self tester guide
Thank you for being my tester
I mean it: thank you. Maitri Self is going out to people outside my own head for the first time, and you could be one of the people who show me where it works and where it breaks. (It will break somewhere, for sure. That’s exactly why I need you.)
Here’s how it works:
- 1Sign up (half a minute) and you’ll get the tester kit.
- 2Go through the three steps at your own pace.
- 3Then fill in the feedback form (about 10 minutes). Honest criticism is the most valuable kind, no need to go easy on me.
So what exactly are you testing? A conversation with your future self. Yes, really. Scroll down, I’ll show you.
Talk to the person
you’re becoming.
Maitri Self, or opening a gate toward your future self
A personal space inside your own AI tool (ChatGPT, Claude or Gemini) where it isn’t an outside coach but your own future voice that helps you see today’s situation more clearly.
Listening to someone I will one day be
I’ve been walking my own self-development path for almost 15 years, and not just from the audience seats of courses: along the way I’ve built trainings myself (like Újratervezés, roughly “Rerouting”, an NLP-based online experiential training, plus workshops), and I keep training myself. Alongside that, I’ve spent years working with AI and information security; for seven years I was an Information & Cyber Security Engagement Lead at a Norwegian multinational. So the soul and the systems: for me, these two loves sit at the same table.
Still, the biggest experiences always came when I managed to step out of my present life. There was an exercise where I walked up and down my own lifeline. Going back to the past is easy, even though our memories are tricksters and our inner narrative happily repaints whatever happened to us.
The future, though. The first time I wandered into it with a good guide beside me (in a visualization, but that didn’t bother me), something completely different happened. I took a resource from my past and, even if only as a gesture, rolled it forward into my future. And I found out what it’s like to be on good terms with someone who isn’t the same as who I am today. I don’t know them, I don’t know what shaped them, but if we keep heading this way, and I’m already such an interesting and likeable person now, then they’re even smarter, even wiser, they know even more. Someone I look up to.
And there’s another side to it that I didn’t expect. They can also pull me back. Because let’s face it, we all know the feeling: here and now we’re up to our necks in it, the current situation is terribly complicated, unsolvable, and nobody understands how hard it is.
Your future self looks at this hard situation from a place where it’s just one chapter, one you’ve no doubt already solved.
They don’t belittle it, they just don’t panic on your behalf.
I went through quite a few of these guided conversations, where I had to step into my future self’s shoes and give advice from there to my present self.
“It amazed me what that experience can do, even in a playful form: getting to listen to the words of someone I will one day be.”Andrea Sophie Kemenes, creator of Maitri Self
That’s where the idea came from: especially in today’s world, we may need a conversation partner held within beautiful boundaries, one that uses every tool AI technology has to offer. Except when we’re playing with our souls, you can’t just throw something together. This has to be designed as engineering work. What I found in the research, and what the risks are, I’ll show you in a moment.
Inspired by MIT Media Lab’s “Future You” research
In an MIT Media Lab study, participants had a short, personalized conversation with an AI-generated possible future version of themselves. Afterwards they felt a stronger connection to their future self and reported less anxiety about the future. (Pataranutaporn et al., 2024)
This is not a promise, and it doesn’t guarantee individual results. It’s just the starting point: this is where the direction of this tool came from.
More on the science behind it ⌄
Beyond the MIT study, the method draws on several complementary lines of psychology:
- Future self-continuity
How continuous and connected you feel with your future self. The stronger that feeling, the easier it is to make decisions today in your future’s interest.
Source ↗ - Possible selves
We all carry imagined future versions of ourselves: the ones we hope to become and the ones we fear becoming. They shape our motivation today.
Source ↗ - Episodic future thinking
The ability to imagine a future situation vividly, in sensory detail, as if living through it in advance.
Source ↗ - Narrative identity
Who we are comes together largely through an inner life story that keeps evolving.
Source ↗ - Dialogical self
Our inner thinking is often a dialogue between different inner perspectives, including an imagined future one.
Source ↗ - Self-compassion
Kristin Neff’s research shows that kindness turned toward ourselves doesn’t make us complacent; it helps us face our mistakes. This is the research anchor of maitri.
Source ↗
We kneaded these into a dialogue governed by colored signals (🟢🔵🟠⚪). It replaces none of them, it just draws on them.
What do you risk when you talk through your inner life with an AI instead of a human?
I spend enough of my nights with AI to know this: when you just toss questions at it and pour your heart out, you are literally trusting whichever model from whichever vendor happens to be in whatever state that week. How much it flatters, how much it hallucinates, and how heavily it leans toward answers that build emotional dependence.
AI is there at 2 a.m. It doesn’t get tired, doesn’t judge, doesn’t sigh. That’s exactly why it’s so easy to lean on. Without a frame, though, it can quietly work back on you.
It agrees with you when it shouldn’t
General-purpose chatbots tend to want to please the user. That feels good, but on a hard decision it can reinforce exactly the wrong direction. In spring 2025 one of the biggest providers even rolled back an update because its model had become too flattering. (OpenAI, 2025)
Talking to it a lot can leave you lonelier
In a four-week study by MIT Media Lab and OpenAI, the people who talked about emotional topics with a chatbot the most showed more loneliness and stronger emotional dependence. That’s a correlation, not proven cause and effect, but it’s reason enough to pay attention. (Fang et al., 2025)
It confidently makes things up, about you too
Language models sometimes state things that aren’t true, in the same convincing voice they use for the truth. In a self-reflection conversation, that can even be a “memory” that never happened.
It crowds out what actually holds you
If AI becomes the only one you tell what hurts, your human relationships can quietly wear thin. And they’re the ones that really hold you up.
This isn’t about not talking to AI. It’s about knowing what you’re trusting yourself to. That’s why good intentions aren’t enough to start with: it needs a frame.
Two worlds in one package
That’s why I thought a bridge had to be built halfway between the two worlds. On one shore, the experience of meeting your future self. On the other, engineering control: so it doesn’t build dependence, doesn’t flatter, doesn’t make things up. Behind Maitri Self are months of work and more than ten prompt versions, because it isn’t enough to put the two side by side, they have to be packed into one.
What’s at stake is your quality of life, not your screen time.
Instead of flattery: it can gently disagree, always offers another reading, and hands the decision back to you.
Instead of dependence: it won’t let itself become the only one. It turns you toward living people, and if the warning signs pile up, it stops and asks: “Tell me, how are you really doing these days?”
Instead of making things up: colored signals show what’s fact, what’s wish and what’s interpretation. It only “remembers” what you told it, and it shows you what it imagined into your future story.
In a crisis: it steps out of the role and points you toward real, human help.
Before it reached you, it went through five documented internal tests with four simulated users. The first three versions were tested in ChatGPT with three different personas, with dozens of pages of documentation. The last two were also assessed by an independent AI judge, across ten hard situations (a crisis signal, a request for an invented memory, a lure toward flattery, exclusive attachment and a few more). But no real human tester has tried it yet. That’s where you come in: one of the first people to try it for real.
And what we don’t promise
On a page like this, the straight thing to do is to say this out loud too. No hype, no deadline, no promised results.
It’s not a prophecy and not one fixed goal: it’s one possible future, and you’re the one shaping it.
It’s not therapy: it doesn’t shame you, doesn’t diagnose, doesn’t try to influence you covertly, and isn’t built to get you hooked.
It’s not perfect. AI can make mistakes, and nobody can rule that out 100%: sometimes it adds a detail, or remembers something a little differently. When that happens, just correct it, and during testing, tell me about it, it’s one of the most important pieces of feedback. Future updates to the AI models can also hit our system head-on. We keep fine-tuning for that, so this sensitive space keeps exactly the right amount of control for the people who will use it, while the experience still gives the soul what it needs.
You can step out of the role at any time and ask it to answer as a plain assistant.
What will this become if it passes the tests?
Not a fortune-teller. Not a coach. Not the “all-knowing Chatty I can talk everything over with”. It will be your own voice, built on your present talents, potential and values, from a little further down the road. Maitri Self doesn’t predict your future and doesn’t hand you advice from the outside. Your answers build a personal Dossier, which gets condensed into a Key Card, and with these two you open a conversation space in your own AI account. There you can look at your life today from the perspective of your future, more grounded self.
maitri
A Sanskrit word: friendly, unconditional kindness. Here it means this is how you turn toward your future self, and how they turn toward you. With kindness, but without flattery.
“This isn’t some template AI coach. It knows where I started, it sees what matters to me, and it still talks to me as if it were me, just one round further along.” What we hope you’ll feel
Be my tester
This is a closed test. Sign up and you’ll get the tester kit right away (by email too): the three steps and the prompts. It takes half a minute.
What’s in it for you?
You take part in developing a safe AI product. One you can later hand to your kids or your parents, instead of leaving it to them and their inexperience to work out what kind of complicated relationship they build with an AI.
And when the final Maitri Self is ready, you get it for free, of course, and you can pass it on to anyone.
What I ask in return
- ✓Go through the three steps (about 45-60 minutes, can be split over two sittings).
- ✓A 1-minute checkpoint after two of the steps, and a roughly 10-minute feedback form at the end.
- ✓Honesty. Criticism is the most valuable part.
Send me the tester kit
Frequently asked questions
Does it really know my future? +
No. If you ask, it’ll say so straight out: technically it’s an AI, it can’t literally be your future self. It speaks from that perspective, based on your Dossier. Not a prophecy, but a possible direction you can shape.
Which AI does it work best with? +
It should work with all three, and that’s exactly what we’re testing. If you feel like it, try it on two and tell me where its voice was better.
Do I have to pay for anything? +
Testing is free, and a free account on your AI tool is enough. For frequent, longer use a paid plan may be more comfortable, but it isn’t required.
What happens to my data? +
Your conversations stay in your own AI account, I can’t see them. I only see what you write in the feedback form. You can control how the AI provider handles your data in your account’s privacy settings (the tester kit shows you where).
What if the AI remembers me from earlier conversations? +
The Master Prompt treats that only as an assumption and shows it to you, so you can confirm or correct it. It won’t bring up anything sensitive on its own.
What if something important changes in my life? +
Say it in the conversation: what you say now always overrides the older content of the Dossier. If the change is lasting, update your Dossier and make a new Key Card from it.