June 2026
Fair question. If you already have access to a capable AI assistant — and most people reading this do — why would you pay for a separate AI coaching service? You can describe your workout to ChatGPT, ask Claude for a training program, or tell Gemini about your shoulder and get a reasonable answer. The quality is often very good. So what's the case for a purpose-built tool?
Part of the answer is about features and coaching quality, which we'll leave for elsewhere. But another part relates to something most people haven't thought through: what happens to the information you share.
When you use a general-purpose AI assistant regularly, you are — over time, across dozens or hundreds of conversations — sharing an extraordinarily detailed picture of yourself. Your work frustrations & career anxieties. Your relationship dynamics. Your financial situation. Your health concerns. Your interests & hobbies. Your political views. Your fears & ambitions.
No single source of information has ever produced this quality of signal about an individual. Data brokers have long assembled profiles from purchase histories, location data, social media activity, and public records. But those profiles are inferred and probabilistic — they know you bought running shoes and visited a cardiologist's website, so they guess at things about you. A conversational AI doesn't guess. You tell it directly, in your own words, with context and nuance, repeatedly over time.
That's what makes these tools so useful. It's also what makes the data they hold so sensitive.
The risk here isn't primarily that AI companies are careless or malicious. Most are neither. The risk is what data scientists call concentration: the more comprehensive a dataset about you, the more valuable it becomes — and the more consequential it is if something goes wrong.
Comprehensive personal profiles are valuable to advertisers, to data brokers who would pay to license insights from them, to governments who can compel disclosure, and to anyone who manages to breach a system. A company can have excellent intentions and still be acquired by a less scrupulous parent, change its privacy policy in response to business pressure, or simply suffer a breach that exposes data it had no malicious intent to misuse.
This isn't hypothetical. Major platforms with strong reputations for security have been breached. Business models that didn't depend on advertising when a product launched have shifted when growth slowed. Terms of service that seemed protective have been rewritten. The risk of concentration isn't about trusting any particular company in the present — it's about anticipating that circumstances will change.
There is a less obvious aspect to this risk that deserves attention, too. Until recently, large collections of unstructured personal data were sensitive in principle but somewhat protected in practice by the sheer effort required to make sense of them. A boatload of chat logs required significant human time and judgment to analyze. That friction was never a formal protection, but it was one that mattered.
AI removes it. The same capabilities that make AI assistants useful — understanding context, finding patterns across unstructured text, connecting information across sources, drawing inferences from what isn't said directly — are precisely the capabilities that make personal data easier to exploit at scale. A dataset that would have taken a team of analysts weeks to mine can now be processed in minutes, at negligible cost, by anyone with access to capable AI tools.
So, the sensitivity of personal data is not fixed. It increases as the tools available to exploit it improve. Information you shared ten years ago, under a privacy policy that seemed reasonable at the time, is now held in a world where the company's ability to extract value from it — or where a “bad actor's” (in the cyber security sense, not the Hollywood sense) ability to misuse it — is meaningfully greater than it was when you agreed to those terms. The data hasn't changed, but the risk calculus associated with that data has.
This is not a reason to avoid AI tools entirely. It is a reason to be thoughtful about which sensitive information goes where — and to recognize that keeping your health and fitness data in a purpose-built service with a transparent business model is a different decision than folding it into a general-purpose platform whose future uses of that data are uncertain.
Health and fitness information sits in a particular category. It is intensely personal, and it has real-world consequences.
Injury history, physical limitations, chronic conditions, and performance data are exactly the kind of information that insurance companies, employers, and marketers find valuable. You may not be sharing this information in a formal medical context — and so it may not carry the legal protections that medical records do — but it is health data in every meaningful sense. Logging your workouts, describing your knee pain, noting that you're managing Achilles tendinopathy: these are disclosures about your physical condition that you probably don't intend to make available to anyone beyond the immediate conversation.
When that information lives inside a general-purpose AI that also knows your employer, your salary anxieties, your family situation, and your political views, you have created a profile of yourself that is comprehensive in ways that would have been impossible to assemble a decade ago. The question of who has access to that profile, under what conditions, and for how long, deserves more attention than most people give it.
Most people have a vague awareness that their data is collected and sold. Few understand the scale of it. There is an entire industry — largely invisible to consumers — whose business is assembling, packaging, and selling personal profiles. These brokers buy data from apps, loyalty programs, and public records; combine it across sources; and sell access to advertisers, insurers, researchers, and anyone else willing to pay.
The profiles they build are probabilistic — assembled from behavioral signals rather than direct disclosure. They know approximately who you are, what you're likely to do, and what you might be persuaded to buy. They are wrong often enough that you'd be unsettled if you saw what they had. They are right often enough to be commercially valuable.
A general AI that holds detailed first-person disclosures from you over years is a fundamentally different kind of dataset — far more accurate, far more nuanced, and far more comprehensive. Whether and how that data might eventually interact with the broader data ecosystem depends on decisions made by companies, regulators, and acquirers that you have no control over.
The simplest defense against that risk is not to put sensitive information in places where it doesn't need to be.
A coaching service designed specifically for fitness, with a business model that doesn't depend on advertising or data monetization, holds a fundamentally different relationship with your information.
Xenos Fit knows what it needs to know to coach you well: your training history, your physical considerations, your goals, your sport. It does not know your salary, your relationship status, or your search history. It has no incentive to share what it does know, because its revenue comes from you directly, not from selling access to you.
This isn't a claim of perfection. Any service that holds personal data carries some risk, and we try to be honest about that on our About page, including what data we collect, why, and how we think about access. What we can say is that the architecture and the business model are aligned with minimizing what we hold and protecting what we do.
When you share health and fitness information with a general AI, you are contributing to a profile that is comprehensive, persistent, and held under terms that may change. When you share the same information with a purpose-built coaching tool, you are using a service that exists specifically to help you train — and whose business depends on you trusting it with exactly that information and nothing more.
The question isn't whether AI is good or bad. The question is whether the information you share to get coaching help needs to be held with everything else you've ever told an AI assistant, or whether it's worth keeping that part of your life separate.
For some people, the convenience of a single AI for everything will outweigh those considerations. That's a legitimate choice. But it should be made with eyes open to the security & privacy tradeoffs.
Further reading