fitness · technology

BMI charts or AI assistants for weight management? Uh, neither, please.

CW: Discussion of weight and weight loss.

AI is the new watchword (well, acronym) in public discussions of pretty much everything from education to finance and now, to health. It’s both lauded and feared for its potential power to replace many of our jobs. But will it live up to that potential? What will be the effects on our work and creative lives?

At the moment, I’ve no idea. But I do have some thoughts about a recent editorial on the “Transformative Role of Artificial Intelligence” in weight management.

Drawing of Woman doing Baby sign language for "WHAT?"
Baby sign language for “WHAT?”

Yeah, I’m with you on this. But let’s hear what the authors have to say here. They start out with acknowledging some of the limitations of the use of BMI (the body mass index chart) for classifying weight. They argue that we need a more “reliable approach to understanding [body weight] and health”. Okay. I’m grudgingly not disagreeing. Yet.

Now enter AI: the solution to all the weight management problems that we probably never really had. AI can, according to the authors, do everything for us. It can:

  • create personalized treatment plans (presumably to reduce weight, based on AI’s view of what we should weigh)
  • develop tailored and more effective intervention strategies (again, for weight loss)
  • predict risk of weight-related medication complications for “a new era of early intervention and preventive healthcare”

Oh really? Color me a bit skeptical here. Somehow, analysis of large data sets of patient biometric information is going to yield revolutionary changes in precision medicine about managing body weight? We haven’t even figured out how to use AI analysis of large data sets to manage resumes and hiring in accurate and non-biased ways, which is a much easier problem.

But wait, the authors have another proposed solution, and it is: AI-driven virtual assistants,

… to provide continuous support and guidance to individuals on their weight-loss journey. These assistants can answer questions, offer motivation, and provide reminders for medication or exercise, enhancing overall patient engagement.

Hmm. Do I want some AI-bot, gobbling up my personal data, sharing/selling it to everyone in sight, telling me what to do and eat and not eat and generally cheering me on to “just do it”?

Gosh, thanks ever so much. But no.

The authors admit that there are some kinks to be worked out in the system, like “ethical considerations, data privacy, and potential biases in algorithms.”

In lieu of an extended rant, I’ll leave it at that. Except to say this: the weight-loss and weight-management industrial complex is currently working itself into a dramatic crescendo of fancy technological and pharmacological approaches to control body weight in the name of health. We have loads of evidence that BMI is a poor tool for classifying and guiding healthcare practitioners in advising and treating patients for better health. We have loads of evidence that AI-driven data analysis for complex decision-making is riven with problems, both methodological and ethical. There’s reason to think we can improve on these methods, but we’re not even close to there yet.

So, in the meantime, keep your AI virtual assistants away from me. Okay?

Kthanksbye

p.s. In the authors’ acknowledgements, they add that they used ChatGPT (GPT-4) in preparing this article. I’ll leave conclusions about that to you, dear readers… Oh, and this blog post? I wrote it all by myself.

Let us know what you think....