Post

The drivers of this algorithm are alcohol use, drug use, and particularly smoking, which is highly correlated with alcohol and drug use. So it seems to be detecting smokers. Which is a first step, but does it add a lot to existing clinical practice?

Blocked that idiot. Thanks for posting the Lancet journal. For me, this is exactly what AI would be beneficial for humanity.

‘Researchers found lifestyle factors were associated with facial ageing. For example, heavy alcohol use, drug use and smoking were linked to looking biologically older in AI analysis of facial photographs’ Do we really need ai to tell us this?

No, but being able to do it from a single photograph is impressive. "Do we really need the internet, we have telephones?" "Do we really need cars, we have horses?" Etc

I’m sure ai can be a v good tool (eg as a diagnostic tool able to analysè lots of data quickly). But this article did not make clear to me what any of them are. It seemed to hint at ai being a useful means for withholding treatment - which may be justifiable, but caution needed, surely?

Internet to telephone comparison falls flat. And yes cars are objectively faster than horses, carry more people and equipment than one horse, etc. No, we don’t need an AI to tell us that known carcinogens cause cancer. That’s like going outside and telling people the sky is blue, like thanks?

I think every single paper ever published has the “while this might be useful, more research is needed to make it more useful or determine how useful it is in practice”

I'm struggling to make sense of the conclusions of this study when the vast majority of people I personally know who have developed cancer are: (i) female (ii) non-smokers (iii) have developed breast cancer (iv) are not in remission or have decided to have double mastectomies. 1/2

Nothing about their 'visage' provides clues as to their vulnerability. Most appeared outwardly healthy during the period of time I've known them until the radiotherapy & chemotherapy commenced.

Different AI. Not LLMs or any GenAI. Just good old fashioned machine learning based on sound mathematical principles. If only the LLM companies weren't hoovering up all the investment money, we'd have more of this.

This is a "breakthrough" algorithm that predicts... mortality increasing with age and unhealthy look. What an advancement of science! Curious that it got even a brief mention in The Guardian though, my expectations were higher.