Category: Physical Activity

Laboratory balance beneath a glowing data network and warning symbol

Activity Monitor, Artificial Intelligence, monitoring, Physical Activity October 26, 2026

Your data, your prescription? What Google’s personal health agent tells us about AI-driven activity prescription

For more than a decade we have been collecting data from wearables at an incredible rate: steps, heart rate, heart rate variability, sleep, GPS traces, training sessions. Yet the step that matters most has always been the hardest one: going from numbers to a decision. What should this person do next week, given their data, their health history and their life? In high-performance sport we solve that problem with teams of people. For the general population, that team simply does not exist.

A paper just published in Nature by Heydari and colleagues at Google Research, “The anatomy of a personal health agent” (21 October 2026), is the most serious attempt I have seen so far to build that team in software. It is worth reading carefully, both for what it shows and for what it does not show, because it gives us a realistic preview of AI-generated, personalised activity prescriptions based on our own data.

What they built

The authors started from users rather than from technology. They analysed 1,370 real consumer health queries from search, Gemini and the Fitbit forum, surveyed 555 people using an experimental Fitbit Labs chatbot and ran a workshop with 14 experts. Four types of need emerged: general health knowledge, interpretation of personal data, wellness advice, and symptom assessment. These needs require very different skills — numerical precision and code for data questions, factual grounding for knowledge and symptoms, empathetic open-ended dialogue for coaching — and the team found that a single large language model struggled to balance them.

Their solution is a modular, multi-agent system — the Personal Health Agent (PHA) — built on Gemini 2.0, made of three specialists coordinated by an orchestrator:

A data science agent that turns a vague question (“am I getting fitter?”) into an explicit analysis plan, then writes and runs Python code on the user’s wearable time series, including comparisons with population percentiles by age and sex.
A domain expert agent that reasons over wearable data, blood tests and health history, using tools such as NCBI literature search, web search and population statistics to ground its answers.
A health coach agent built around motivational interviewing: it identifies goals first, explores constraints, asks for feedback and only then offers a SMART recommendation.

The orchestrator decides which agent leads and which ones support, rewrites a targeted sub-question for each, checks the combined answer for coherence (and to avoid asking the user for data it already has), and keeps a memory of goals, barriers and preferences across turns.

What the evaluation shows

The evaluation is large: 11 benchmark tasks, more than 7,000 annotations and about 1,350 hours of work from end users and experts, grounded in the WEAR-ME study (1,165 Fitbit users with blood tests and questionnaires). A few results stood out for me.

Data analysis. The data science agent’s analysis plans scored 75.6% against 53.7% for the base model. The most striking gain was in checking whether there was enough data to answer the question in the first place: 66.5% of plans did so, against just 5.3% for the base model. Anyone who has worked with wearable data knows that missing days, non-wear periods and artefacts are where most bad conclusions come from. Its code passed 75.5% of expert-written unit tests at the first attempt (58.4% for the base model), rising to 79% with self-correction.

Health knowledge and context. On more than 2,000 board-style questions in endocrinology, cardiology, sleep medicine and strength and conditioning, the gain over the base model was small (83.6% vs 81.8%). Much bigger differences appeared when personal context mattered: end users rated 96.9% of the domain expert’s personalised answers as trustworthy against 38.7% for the base model, and five clinicians judged 65% of its multimodal health summaries useful against 30%.

Coaching. Users preferred the coach agent on five of six coaching themes, and expert coaches preferred it on all six. Interestingly, both groups gave the plain baseline a slight edge on progress measurement — a weakness the authors then addressed in the integrated system.

The integrated agent. In live multi-turn conversations, 45 users role-playing real WEAR-ME profiles ranked the orchestrated PHA first 46.7% of the time, against 28.9% for a parallel multi-agent design and 24.4% for a single agent. The experts were far more decisive: the PHA was ranked best in 66.7% of conversations, roughly five times as often as the single agent, with an overall quality score of 72.8% against 64.3% and 54.4%.

Why this matters for activity prescription

What struck me most is that the architecture mirrors how a good performance support team works. Someone looks at the data properly, someone interprets it in the context of the person’s health and history, and someone turns it into a plan the person will actually follow. Exercise prescription needs all three. Frequency, intensity, time and type are not chosen from a guideline table alone; they depend on where the person is starting from, what they can safely tolerate, and what they are willing and able to do.

Seen through that lens, the near-future pipeline is easy to imagine:

Baseline from your own data: real habitual steps, active minutes, resting heart rate, HRV and sleep — computed, not guessed, and checked for missing data.
Comparison and context: where you sit relative to people of your age and sex, and what your blood markers and conditions mean for what is safe.
A negotiated dose: a first step that fits your constraints and motivation rather than a generic “150 minutes a week”.
Progression: memory of what was agreed and what actually happened, so the next prescription builds on the last one.

The example conversation in the paper’s extended data shows this already. A user with type 2 diabetes, a BMI of about 36, around 5,500 steps per day and a recent hospital admission asks how to get back to 10,000 steps. Instead of prescribing the target, the agent explores why it matters to her, picks up symptoms that suggest possible sleep apnoea, recommends speaking to her doctor, and agrees on several short 10-minute walks per day as a starting point. That is, in essence, a graded prescription with a red-flag screen built in. Another example I liked: asked whether more activity leads to more deep sleep, the data science agent reported a correlation of about 0.14 and said plainly that the relationship was weak — exactly the kind of honest answer a good analyst gives and a lot of consumer apps do not.

The real opportunity is scale. One practitioner cannot review thousands of people’s data every day and adjust their plan. A system like this could make the first layer of individualised prescription and follow-up available to anyone wearing a device, and free human experts to focus on the people and decisions that need them.

The risks we should not gloss over

The authors are refreshingly upfront about limitations, and they describe the PHA as a research prototype rather than a clinical tool. Still, for anyone tempted to treat this as a ready-made prescription engine, these are the issues I would keep in mind.

Safety at scale. In the expert safety review of the clinical summaries, 83% carried no or minimal potential for harm, but 15% were judged mildly harmful and 2% moderately to severely harmful. Two per cent sounds small until it is multiplied by millions of users.
No outcome data. Everything measured is conversation quality and short-term task performance. There is no evidence yet that people change behaviour, get fitter or become healthier. For an activity prescription, the only test that ultimately matters is the dose–response over weeks and months.
Data quality and generalisability. The data come from one device ecosystem (Fitbit), and consumer estimates of sleep stages or HRV have their own validity limits. Even the improved agent checked data sufficiency in only two-thirds of its plans, so one-third did not.
Simulated users. In the live studies, participants role-played personas matched to their own profiles rather than seeking help for themselves, and parts of the evaluation relied on LLM-based autoraters. Both are reasonable research choices, but neither is the real world.
Training load is not in scope. The system was not evaluated on periodisation, load management, injury risk, environmental stress such as heat, or athletic populations. Extending it to performance settings needs its own validation.
Evidence quality. Open-web retrieval can surface weak or conflicting evidence. The authors themselves suggest restricting retrieval to curated sources and showing provenance next to every recommendation.
Privacy, equity and cost. Real deployment needs proper privacy and security controls, fairness checks across populations, languages and literacy levels, and a cheaper architecture: the PHA averaged 6.5 model calls and about 205 seconds per response, against 3 calls and 36 seconds for a single agent.
Who built and judged it. The work was funded by Alphabet, most authors are Google employees, and the system runs on Google models and Fitbit data. That does not invalidate the findings — and releasing the code and benchmark is a genuinely good step — but independent replication will matter.
Coaching versus prescribing. The coach agent sometimes declines to give a concrete number and steers the user to decide for themselves. Good motivational practice, but people asking for an exercise prescription often need a clear, specific dose — and someone accountable for it.

What I would want to see next

Before trusting an AI-generated activity prescription, I would want to know four things: whether it checked the quality of the data before drawing conclusions, whether its evidence comes from curated sources it shows me, whether it recognises red flags and hands off to a clinician rather than reassuring, and whether its prescriptions are reviewed against what actually changed over time. The next generation of studies should be prospective, longitudinal and outcome-based, with people using the system for themselves, across devices and populations, and with clinicians and coaches in the loop.

Bottom line

This paper does not show that AI can prescribe exercise better than a qualified professional. It shows something more modest and, I think, more important: that splitting the job into data analysis, domain expertise and coaching, and orchestrating them carefully, produces answers that experts judge to be substantially better than a single general model. Personalised activity prescriptions built from our own data are no longer a distant idea. The challenge now is to prove that they are safe and that they actually change behaviour and health — and to make sure the humans who are accountable for those decisions stay firmly in the loop.

Reference: Heydari, A. A. et al. The anatomy of a personal health agent. Nature (2026). https://doi.org/10.1038/s41586-026-11161-8 (open access). Code: Google-Health/consumer-health-research.

Athletes, Exercise, Physical Activity December 6, 2011February 9, 2013

You cannot be serious!

I have a day off today. I am supposed to chill out and try to avoid work at all costs. There will be few of such days before we get to the Olympics in London so I should make the most of it.

When I am off work, I read. This is what i like to do, reading and learning. I could read for hours about anything and everything. I never get bored.

Today I decided to scan the press first. My fault. I grew up thinking that journalism was about reporting the truth and/or writing evidence-based opinions. We all know this is not the case anymore. Despite sparse moments of well written brilliance, we are now inundated by a sea of absolute and utter bullshit. As well as non-sense involving celebrities being famous for God knows what.

In the sea of badly written nonsense, I came across this one written by Liz Jones on the Daily Mail. Liz is a British Journalist and writer currently writing columns for the Daily Mail, The Mail on Sunday and British Airways’ High Life Magazine.

Liz is a woman of style. In fact she writes about fashion mostly. This time her column is about sport and physical activity. Her piece on the Daily Mail is entitled: “Sinewy arms are NOT a good thing – and neither is sport”. It is well written and has style, just like every piece Liz has written. There is only one problem with it: it is so badly wrong and full of nonsense that deserves some serious comments. So serious that I decided to write about this on my blog rather than spending my day off with a good book in hand. It actually deserves a letter, but letters to editors or newspapers most of the times end up in bins, and online comments get lost. So my letter will be on this blog.

I suggest the readers of my blog to read the article and make their mind up. This article on my blog is a reflection of my opinion and I may be wrong, but I decided to put it here anyway. Liz’s article pretty much states that “athletes could have spent their precious time more productively reading or studying maths or helping people rather than exercising” or “ Sport in school is the worst thing you can possibly inflict on children”.

So, here it is.

Dear Liz,

I came across your article on the Daily Mail online entitled “Sinewy arms are NOT a good thing – and neither is sport”. One of the coaches I work with shared the link on twitter (so it’s his fault really!) and I could not resist spending some time reading it. I read it with interest, just like everything I read. However, after few lines of hope, I was hit with your statements and opinions. I am not going to comment on your views about one of our greatest athletes “not being seen as a great advertisement for young women who might be thinking about taking up a sport”. Simply because you totally missed the point.

Sport is about hard work and sacrifice. I grew up being taught that hard work and sacrifice will always make you a better person. I believe in it, this is what I try to teach my son. Physical exercise and sport are about hard work and commitment. Healthy diets are about commitment and hard work. So when someone develops the “sinewy arm” as you call it, after hard work, sacrifice and commitment, without the help of illegal and harmful drugs should not been condemned about being a bad advertisement for young people. It should actually be considered a positive model.

The above is not what worries me, Liz. It is the rest of the article that worries me. An intelligent woman like you, which clearly spent a lot of time reading, studying maths or helping people instead of doing sport should not have been overcome by emotions when writing the article. You seemed to have had quite a sad experience with school sport. I read about your terrible experiences in running in a sea of mud while playing hockey in the winter or your vomit-inducing cross country efforts. Not something I recommend to anyone. What worries me is you being humiliated in public when running cross country. I am not sure about the quality of the physical education programme in Essex at the time you were at school. But I am not aware of public humiliation being part of a cross-country competition. Also, the vomiting bit is not a mandatory experience in a cross country race. It is something which can happen for a variety of reasons, such as stress, wrong timing and content of a meal and lack of training as well as trying to run too fast. I am sure it was a bad experience, but it was your own. I am sure some appropriate advice on what to eat and when as well as more regular exercise would have made your experience a lot better.

Swimming was outside in the cold, gym on thin rubber mats, and in Netball and you were the small one pushed aside by the “bigger girls”. It was tough for sure. It sounds in fact so bad, I urge you to tell us what School you are talking about. It seems to me that a school with such poor standards in physical education and sport could have done some real damage also to your reading and doing maths experience. Your words reinforce even more what I think. I think there should be more money going into school sport. More money to make sure the swimming pools are warm, hockey is played on astroturf instead of mud and women get more choice than just netball (maybe handball, volleyball and basketball if they are into indoor team sports). Most of all, more money should be invested in improving the PE curriculum, the quality of teachers and the possibilities for every child to practice the sport of choice.

Your avoidance of swimming lessons was a choice. A choice you and your mum made. A choice which has not given you the opportunity to appreciate the joys of swimming and the ability to do it safely.

Condemning school sport and physical activity is dangerous and wrong. Your bad experiences don’t necessarily mean that sport in school is everywhere as bad as yours was. Most of all you should know what the consequences of unhealthy sedentary lifestyle are in order to understand better what could be the consequences of what you write.

A woman of knowledge like you should have spent some time doing research. You had been spending a lot of time reading at school after all. Hence, you should know where to find the information. Let me help you. Here are some of the stats about obesity and physical activity published by the NHS.

Using both BMI and waist circumference to assess risk of health problems, for men: 20% were estimated to be at increased risk; 14% at high risk and 21% at very high risk in 2008. Equivalent figures for women were: 15% at increased risk; 17% at high risk and 24% at very high risk.

In 2008, 16.8% of boys aged 2 to 15, and 15.2% of girls were classed as obese, an increase from 11.1% and 12.2% respectively in 1995. Whilst there have been marked increases in the prevalence of obesity since 1995, the prevalence of overweight children aged 2 to 15 has remained largely unchanged (values were 14.6% in boys and 14.0% in girls in 2008).

For boys, on weekdays, the proportion who spent 4 or more hours doing sedentary activities was 35% of those who were not overweight or obese, 44% of those classed as overweight and 47% of those classed as obese in 2008. For girls, a comparable pattern was found; 37%, 43% and 51% respectively.
In 2008, boys aged 2 to 15 were more likely than girls to meet the recommended levels of physical activity with 32% of boys and 24% of girls reporting taking part in 60 minutes or more of physical activity on each of the seven days in the previous week.

Almost two thirds of children who had attended school, nursery or playgroup in the last week had walked to or from school on at least one day in the last week (63% of boys and 65% of girls) in 2008.

Among boys aged 2 to10, more met the physical activity recommendations for children if their parents did so for adults. Among girls, the activity level of parents made relatively little difference to the proportion meeting recommendations, but those who had parents with low activity levels were considerably more likely to be in the low activity category themselves.

If you cannot be bothered reading the NHS data, read the summary above and the last point very carefully.

As you can see Liz, kids are getting bigger and bigger. And they are more likely to get overweight if their parents don’t do any physical activity. This is not bad news for how they will look like. It is incredibly bad news for their future health and the taxes you and me pay. If you are interested in serious science (the one you can find in a University library) you should look at this work published on Lancet (one of the most prestigious medical journals). But I am a generous man. I will do the work for you. Here is what you need to know: “the combined medical costs associated with the treatment of obesity are estimated to increase by £1·9-2 billion/year in the UK by 2030″.

That’s a lot of money Liz. Money that me and you will have to generate paying taxes.

I like to pay taxes to improve the quality of life, not to fix errors made in policies and choices. I would like to think that the money I pay provides better libraries, better schools, better roads, safer environments for kids to express their talents and better sports and exercise facilities and coaching. Better medicine, better access to arts. Are you keen to spend your taxes on kids that did not do any physical activity at school, Liz? Are you seriously happy for young people not to understand the importance of a healthy lifestyle, physical activity and diet?

Liz, sport, exercise and diet are all part of a healthy lifestyle you should learn in school. Just like reading and maths. Schools should educate young people not only about using one organ of their body (the brain) but the rest of it as well. We humans, in fact have bodies. This is what we use every day. We use our bodies to move, communicate, work and express ourselves. Bodies are not only a means of transport for the brains. We need to take care of them in order to have longevity and quality of life. Sport and exercise don’t do good only to your body. They do a lot of good to your brain as well, in particular when combined with healthy diets.

Schools and parents can accomplish amazing things if they work together. A recent paper published by Swedish scientists has shown that it is possible to promote a healthy lifestyle and a normal weight development among children from low-income districts with relatively limited efforts involving parents.

Unfit kids are bad news not only for their bodies. A recent study published on Medicine and Science in Sports and Exercise suggested that lower-fit children showed poorer recognition memory performance than higher-fit children in some memory tasks. And you should also read a recent review which clearly summarises currently published research in stating that moderate physical activity is important for youth whose brains are highly plastic and perhaps even more critical for young people with physical disability. Or this one which suggests that evidence shows small but consistent associations between sedentary screen time and poorer mental health. Physical activity and sport also have a positive influence on self-esteem, Liz (again for a scientific review, see here). So, what you wrote about avoiding sport is dangerous and wrong.

I have a kid Liz, and I care. This is why I am writing this. When I leave him at the school door every morning, I hope his teachers will ignite his passion for many things. I hope he enjoys every minute he is using his brain and his body and I hope he develops as a confident, competent and most of all healthy individual. I hope he competes in sport to learn about winning and losing, to learn about respect and hard work, to admire other people’s abilities and develop his own to his maximum, to make friends, to travel and see the world and to meet other people which share his passions. This is what sport gave me and I never got anywhere near winning an Olympic medal as an athlete.

In the last part of your article you then state that today you have to be huge, focused, boring and sponsored to the hilt. I am not sure how you can get it so badly wrong. Allow me to help you again and direct you towards useful information. Your reading abilities will come handy when learning about the athletes training every day with no financial support or sponsors in many Olympic sports (here, here, here, here).

Finally, you state that “The Government and Seb Coe might want us all to run around of a Sunday morning, freezing, but these extremes of excellence are not good for us, surely?”.

Liz, the extremes of excellence give us the inspiration we need in every field. I am sure at some point you must have admired an excellent writer. Someone able to win the Pulitzer prize.

I am not sure where your passion lies. I hope you like arts. Just like the extreme of excellence of Picasso or Leonardo Da Vinci tell us that we will never be as good as them in painting, we should still enjoy doing it and be inspired by them as well as admiring what they accomplished.

In sport, in this country, we are blessed with talent. People who work hard every single day to reach excellence. Excellence which inspires and will inspire our children to work hard every day to be best they can be in every field and in everything they do. So, Liz, if you don’t fancy running in the cold mornings (I don’t do it either!), there are lots of other things you can do which will do a lot of good to your health.

However, if you can’t be bothered to exercise or have problems with sports, don’t demonise it, as our health bill is soaring already and we don’t need further incentives.

Lastly, it is probably worthwhile letting you know that many athletes don’t only exercise a lot their bodies. They exercise their brains too and do reading, maths and help people. If you are short of examples, read something about Katherine Grainger and Tim Brabants. You will find out that their “sinewy” arms have been also used for studying and helping people.