revtothelimit - motorbike forum

Anthropic just mapped out which jobs AI could potentially replace. A ‘Great Recession for white-collar workers’ is absolutely possible

Anthropic researchers found that actual AI adoption is a fraction of what AI tools are capable of performing.

Anthropic, the artificial intelligence (AI) company that emerged in 2026 as an existential threat to billions of market value, with each breathtaking new capability from its Claude model, is back with a warning about just how obsolete AI tools could make whole swathes of work. The AI giant, founded by former OpenAI workers who were obsessed with AI safety just as much as advancement, has been a thought leader on AI risk as much as advancement, and just published a study with the most detailed map yet of which jobs AI is actively performing versus which it merely could perform. The gap between those two numbers is both reassuring and alarming, depending on your line of work.



From:
https://fortune.com/2026/03/06/ai-job-l ... r-workers/

Possibly, probably, but I'll keep arguing against it until it's actually as reliable as a fully trained human with decades of experience.

I thought it would be able to do more in the transport area than that graph suggests.

Replace - no

Significantly simplify and speed up a very information heavy, complex, long, drawn out process, I bloody hope so because it can take years.

Rockburner wrote: Sun Mar 08, 2026 5:16 pm Possibly, probably, but I'll keep arguing against it until it's actually as reliable as a fully trained human with decades of experience.

OK. Moments with Google ...

https://www.bbc.co.uk/news/articles/ckdpg5p820xo

Artificial Intelligence (AI) can analyse X-rays and diagnose medical issues just as well as doctors, a study has claimed.

Software was trained using chest X-rays from more than 1.5m patients, and scanned for 37 possible conditions.

It was just as accurate or more accurate than doctors' analysis at the time the image was taken for 35 out of 37 conditions, the University of Warwick said.


https://oxfordbrc.nihr.ac.uk/ai-analysi ... n-advance/

A University of Oxford research team has developed an artificial intelligence (AI) technology which can accurately predict the risk of a heart attack, heart failure or cardiac death from routine cardiac CT scans, up to ten years in advance.


https://digital.nhs.uk/services/ai-know ... m-ct-scans

George Eliot Hospital approached the NHS AI Lab Skunkworks team with an idea to use AI to speed up the analysis of computerised tomography (CT) scans.

Outcomes and lessons learned

The hoped-for benefits of the project have had some partial successes:

Scan alignment
The team achieved both rigid and non-rigid 3D alignment that were an improvement on manual alignment but not perfect. The methods used to deal with difficulties caused by patients inhaling or exhaling during scanning were observed to work in most cases.

Overlay
Precise overlay of scans was achieved in the tool created both for 3D and 2D images, including when zooming, rotating or panning the image.

New tissue growth detection
Anomaly sizes were measured in 3D but more work is needed for robust correspondence between existing lesions to measure change. 3D to successfully aid the identification of new growths.

Time saving
Further development and testing is required to establish a reduction in radiologists’ time, but the tool provides a process that is less manual for radiologists.

Wasn't there something not 'that' long ago where a computer (AI ?) was shown photos of cancer and non cancer growths and supposedly learnt to recognise which were which - until tested and it didn't. At which point someone realised that for the teaching bit, all the cancer photos had a specific marking or a measure or something. So when shown a photo without that, the AI thought it was benign!!!


I know it's only as smart as the input, but have you met most humans???

Most humans? Ahem ...

https://www.bbc.co.uk/news/health-21466529

There is something odd about this scan of a patient's lung. Have you spotted it yet? How about the dancing gorilla on the right?

I had to go to an IT department work Town Hall. Croissants, coffee, talking, presentations, team table games etc and after lunch we played a game at the back of the room, similar to the old TV program Runaround, where they taped a line down the centre of the floor, asked a question and you had to stand on either side of the room for "yes" or "no".

They asked "Who thinks AI will take their job?". To be honest, I don't know and don't care as I'd be in for a good redundancy package. The whole room went to the "no" side and there was no room left for me to join them so I stood alone on the "yes" side thinking "I can blag this one if I have to". I was given assurance that my job was safe. Damn!

I would suspect everyone …
Unless you are a highly skilled pastry chef at the Dorchester ….

Noggin wrote: Sun Mar 08, 2026 7:01 pm
At which point someone realised that for the teaching bit, all the cancer photos had a specific marking or a measure or something. So when shown a photo without that, the AI thought it was benign!!!

I know it's only as smart as the input ...

GIGO - garbage in, garbage out.

However:

. .
From:
.
.

One of the really frightening issues is that AI is going to be used to rewrite history. Whilst most of us are too old for it to really affect us, it will have profound effect on the next generations and their concepts of right and wrong

I doubt AI will replace my current job in my lifetime (or working lifetime) - the reports of the self drive vehicles don't seem that positive. Maybe possibly in cities or towns with dedicated lanes but up and down twisty and snowy mountain roads ???

'Transportation' has second lowest rating on the chart, after grounds maintenance*

* Even there, robot mowers are being used on golf courses

It's already happening. It will kill a lot of jobs. Some industries will to some extent completely collapse .Knock on impacts to investment in infrastructure and housing, will probably collapse higher education, and reduction in jobs leading to reduction in amounts of taxation collectable.

It's a doomsday in some respects but nothing to do with machines becoming sentient and blowing us up - it's machines just being smarmily better at a lot of those post tech-revolution jobs that didn't exist 75 years ago. And it's done it in 18 months, not over a generation or so.

DefTrap wrote: Mon Mar 09, 2026 10:01 am .... education, and ... it's machines just being smarmily better at a lot

Filly was intrigued that 'education' didn't get a good 'safe' score.

She worked in the field for 30+ years, much of it in classrooms. Many hours spent planning lessons.

So I asked her to open Chat GPT and type in:

"I am a UK key stage 2 teacher. Create a bullet point list of the main elements needed for SATS maths."

Seconds later, she was reading something that would have taken some work to do.

It even included outline details of activities, etc.

Horse wrote: Mon Mar 09, 2026 10:14 am

DefTrap wrote: Mon Mar 09, 2026 10:01 am .... education, and ... it's machines just being smarmily better at a lot

Filly was intrigued that 'education' didn't get a good 'safe' score.

She worked in the field for 30+ years, much of it in classrooms. Many hours spent planning lessons.

So I asked her to open Chat GPT and type in:

"I am a UK key stage 2 teacher. Create a bullet point list of the main elements needed for SATS maths."

Seconds later, she was reading something that would have taken some work to do.

It even included outline details of activities, etc.

It would be nice to think that it will lighten the admin (and 'out of hours') load, releasing teachers to teach and not just be used to reduce the headcount.

I won't hold my breath waiting or bet on it.

I look forward to it helping me! It should increase my studio throughput and make my life easier! It can't replace my role though, as I manually handle rare, fragile and high value items.

Horse wrote: Mon Mar 09, 2026 10:14 am Filly was intrigued that 'education' didn't get a good 'safe' score.

There's an awful lot of folk who are going to get a terrible shock when they realise this. On one hand it's fantastic - who doesn't want to get a head start on (using your example) a lesson plan that might have previously taken hours? On the plus-side maybe it gives you the bandwidth to do other lesson preparation?. This stage one - seems great - but it got here in about 18 months.

Stage two - maybe A.I takes the jobs of support staff doing some of the more humdrum tasks. Probably if A.I is generating ALL the lesson plans then all you need is (fewer) warm bodies to present the A.I-generated material.

Stage three - the next step is a bot on a screen taking lessons.

Given the speed of the tech. advancement and the rush of all sorts of industries not to get left behind, this can happen frighteningly quickly.
People will say "yeah but I do this <example>" as if that means they are protected. Nobody cares - if it cheaper or quicker or it increases the industry competitive advantage, then you're gone.

I think a lot of people don't realise that advances in AI are flying in at an incredible rate.
Looking at things that AI failed at, even as recently as a year ago, are largely no longer relevant as those tools have been superceded.

I work in a very AI-adjacent/overlapping space and the stuff we can do now is frankly insane.

I see it as a bank of 'efficiency technologies'. It's not all ChatGPT, though Claude is bastard excellent.

Currently, I'm seeing that experienced minds can still add value on top of its outputs in a range of areas, but I'm concerned for young people entering industries where their outputs will be less valuable and more expensive than what AI can churn out. It'll drive a skills gap rapidly.