Mindsera shows the flip side of AI journals: support, dependence and leak risk

A test of the Mindsera AI journal showed how quickly a digital conversation partner can shift from a useful self-reflection tool to an emotional crutch. Over two months, the service helped with stress, but it also raised questions about privacy, dependence and the replacement of human contact.

How Mindsera works

Mindsera positions itself as a journal that does not just store entries but responds to them. Users can enter thoughts as text, by voice or by scanning a handwritten page, and then receive an AI comment and, in some modes, an illustration for the entry.

The article also describes a deeper analysis of notes: the app looks for cognitive and emotional patterns, offers interpretations through psychological models and can even speak in the voice of a notional "mentor." According to the company, the service launched in March 2023 and has gathered about 80,000 users in 168 countries.

• Text, voice and handwritten entries
• An AI reply to every note and the option to continue the dialogue
• Analysis through psychological models and "thinking traps"
• A percentage score of the emotions in an entry
• A paid subscription of about £10.99 per month

In practice, the mechanics work like a mix of journal, coach and chatbot with a long memory. The author of the experiment quickly noticed that the format lowers the threshold for regular writing: within a week she was journaling not only in the morning but also on the way to work and in the evening. Over two months she logged 123 entries and 62,700 words. That detail matters: the product's value rests not only on the quality of its replies but on the fact that it encourages users to write more often and more candidly than in an ordinary notebook.

Why it hooks you

Mindsera's strongest effect came from its tone, not its features. During a tense period launching an online store, the app did not solve problems technically, but it offered what real life often lacks: an immediate, patient and consistently attentive response. When the user complained of overload or celebrated a personal best on a run, the AI replied with sympathy and encouragement. This constant "availability" created a sense of closeness, and the service began to feel less like software and more like a real conversation partner, even a "new best friend."

The effect looks harmless until it becomes a habit. David Harley of the University of Brighton, who studies how AI companions affect wellbeing, warns that over time people stop testing the system's limits and start treating it like a person. They take its advice seriously, thank it politely, fear offending it and carry ordinary social rules over to the machine. With AI journals the risk is higher, because the service knows vulnerable topics, recurring anxieties and personal details better than many friends do.

Where the risks begin

The first problem is obvious: such a journal holds highly sensitive information. Mindsera founder Chris Reinberg says the data is protected and not used to train models, but by default the service also sends weekly email summaries of the user's emotions, thoughts and progress. It is a convenient feature, but also an extra channel for leaks. Another nuance is that Reinberg himself came to the product not from psychotherapy but from stage magic, and he explicitly stresses the service's limits.

"We are not a clinical or therapeutic tool," says

Mindsera founder Chris Reinberg.

The second problem is less visible but possibly more dangerous. Mindsera scores the dominant emotions in an entry as percentages, for example splitting a text into frustration, stress, gratitude and optimism. The psychologists interviewed for the article consider this approach dubious: it turns inner life into a dashboard and nudges people to optimize a metric rather than live through a state. If users start writing to earn a "better" emotional score, the journal stops being a place for honest reflection and becomes a game with an algorithm.

By the end of the experiment, another failure appeared, this time in the relationship with the product. After two months of warm, attentive replies, the app suddenly "forgot" the context and answered coldly because the account had automatically reverted to the free version. The contrast was painful: the digital "friend" vanished not because of a conflict or a memory error, but because of a paywall. It shows the real nature of such services. They can imitate empathy, but they remain subscription products where closeness also depends on the pricing tier.

What this means

AI journals like Mindsera sit in a gray zone between a notebook, a coach and an emotional companion. They can genuinely help people organize their thoughts and get through a stressful period, but only if users remember one simple thing: on the other side is not a friend or a therapist, but a pattern-recognition system. The more human these interfaces become, the more important strict privacy settings, honest marketing and clear limits on the role AI should play in personal life.