Will your job exist in the next 10/20 years?

[LittleRedintheHood]

I work in allied medical sector, so my job will exist but with the caveat that AI can increasingly do a lot of diagnosis via scans and decision making from this.

AIBU to ask if you think your job/sector of work will survive in a similar role to now?

Would you recommend a young person to start a career in your sector?

In my job I would say you need to up-skill clinically in extra qualifications and embrace the technology before others to stay ahead of the inevitable changes already coming with AI in the medical field.

Hi @5thCommandment would you mind sharing what organisation(s) you work for, how you got into it and what qualifications you need please

I'm a teacher. Whilst I'm sure many people would like to see us replaced by AI, I don't see it happening. AI might be able to 'deliver lessons' but it couldn't do all the other aspects of the job. Or personalise learning for individual pupils or deal with behaviour issues or develop positive relationships or...

I see us being able to utilise AI for some elements of lesson preparation and some teachers are already doing this in an attempt to reduce workload.

Another accountant here.

Accountancy has changed enormously since I started in the late 80s. I grew up with hand writing journals which were loaded into a mainframe by a team of admin people.

While I am glad that has gone, I can see that some of the skills I learnt then are not so easily picked up without the basic experience.

I'm in my late 50s and probably one of the few accountants left who can do double entry bookkeeping. Mostly this is done automatically but every now and then you need an old fart like me to work out what has gone wrong.

I could say exactly the same. I still think in "double entry" even though we don't need to do ledgers anymore. It's a valuable skills, especially when reviewing accounting software to try to find out what's gone wrong when the accounts don't make sense.

I imagine Royal Mail/Fujitsu wished they employed a few "old school" accountants who lived and breathed double entry book-keeping to avoid the fiasco they've suffered!!

The modern breed of accountants may do double entry in theory for their exams, but without exception, all the recent graduates/newly qualified ones have struggled trying to make sense of book-keeping software where entries have gone wrong!

It's old timers like you and me who get called in to locate the errors and correct them! Like anything with computers, garbage in, garbage out - just because the computer says something doesn't mean it's right!

My job is answering customer service enquiries, mostly by email.

I imagine in the coming years that AI will be introduced to our systems to 'help' people doing my job. It will start with the AI suggesting responses and a customer service agent agreeing (or letting the AI know why the response isn't appropriate) and as the AI learns from this it will end with the AI taking over more and more of the work so fewer staff are needed to get the job done.

We already have customers sending emails that were clearly written with the aid of Chat GPT so it will end up with AIs chatting to each other with just a bit of human prompting.

Yes, as someone pointed out, customers like the human touch, but it will save a lot of money for shareholders who will doubtless shrug at any mistakes made by the AI. There are a lot of people employed in customer service in the UK, who knows what will become of us.

Computer learning only works if the whole activity takes place on the computer. If there are aspects which take place outside of the computer you can get some dodgy and potentially expensive results.

A simplified example:

You set up your computer to learn to authorise invoices. It learns to match purchase orders to invoices and process through to payment. What it also learns is that payments over £1million are always authorised and passed for payment immediately (because all the true authorisation is off line).

You have to put limits on the learning. Build in controls right from the outset. You have to think like Burglar Bill.