Topic: Scientists v AI

[reddebs]

I know we have a few scientists on the forum so I wanted to know how they view their roles with the emergence of AI.

Is it possible or even feasible that scientific researchers for example could be replaced by robots?  I understand that maybe the data analysis could be but I'm not entirely sure.

My reasons for asking is my eldest granddaughter now 13, loves science especially bugs and insects which I'd like to encourage. 

From the research I've done into career options it seems that entomology is a very specialist field especially in the food and agriculture  industries but is it likely to be one that's taken over by AI due to the scarcity of scientists?

[redbyrdz]

Nah, AI is likely to help with research, not replace researchers. It is going to be (and already is in some fields) another tool, not something that replaces the person using the tools.

For example today I used an AI program to help identify things in an image. At the moment, that is still a process that needs a lot of work by a human - I'm doing a pre-selection of the images, I need to manually mark some of the things I am looking for and train the program to find them - meaning I need to check some results, mark some other features, remove some, etc, until the program does a decent job of identifying stuff. But then I can give it thousands of images and it'll spit out the things it found, much quicker than me looking at all of them myself (I still check a few to make sure its correct). In a few years, I can probably do the whole training process much quicker, and I can maybe be more vague about it. But I will still need to decide what to look for in which images, and decide something based on the results.

For example, if I were to look at garden pests, I could imagine that soon you will be able to go and look at a plant, and if you find some sort of bug, take out your phone and take an image, and it'll tell you what it is, and maybe also if you need to do something about it, and what. You might even imagine there could be some sort of robot crawling through a greehouse, take pictures, identify pests, and automatically adjust watering, temperature, pesticides etc. But that robot will replace the manual gardeners, not the pest researchers. Someone needs to tell that robot how to deal with each pest, and check that the result is what is wanted. For example, make sure that the yield doesn't reduce, or does one measure to combat certain pests doesn't bring in others.  (AI can likely also help with understanding how pests are connected, but it will again need direction as to what to look at).


I did talk to a few insect researchers recently, and it is amazing what we don't know about them. There are a few that has been researched very well (fruit flies, for example), but loads we barely know the basics. If she is interested in insects, there'll be loads of work for her.

[reddebs]

^^^  Thanks for all that.

Yeah she loves her bugs so I've just got to hope that school doesn't dampen that enthusiasm as she frequently complains that science is boring at school.

Ooh quick question then, is it absolutely necessary to do a PhD?

[thaddeus]

I am dragging my way slowly through a part-time PhD in Data Science.  In my opinion AI, if paired with robotics, will ultimately have the potential to replace almost every job so I wouldn't base career advice on whether a particular career will become obsolete.

My expectation is that automation of jobs will be profit driven.  BT is a good example as they think can replace their call handlers with a chatbots that do a passable enough job to tick the box of customer support.  I don't really see the profit motive in automating entomology.

As redbyrdz said, in most careers AI will assist rather than replace.  It will largely replace the mundane and the repetitive parts of most jobs if the company or person has the motivation and skills to implement it.

The problem is the jobs that are almost entirely mundane and repetitive and the lack of sufficient gainful employment opportunities elsewhere.  Of course, we could all just reduce to 30 hour working weeks, then 20, then 10... as the spoils of humanity's innovation are spread evenly amongst us all 

Edit: in response to your edit I would say Yes that a PhD in almost every scientific field is necessary now.  My wife has a Masters in a similar field but now they only recruit people with a relevant PhD to do what are largely entry level jobs.  There is always the path less trodden to establish a career in scientific fields without a huge weight of academia behind you but it's more difficult.  As someone not particularly enjoying the PhD experience I would say to anyone to only do it if it's a passion and not simply for career enhancement!

[reddebs]

^^^ again thanks for that.

Interesting about the necessity for the PhD but it's such a pity there isn't a more practical route for the less academic individuals.

She's only 13 so probably doesn't even know there are career options with bugs.  I'll have a good chat with her on the drive here on Friday, let the idea trickle in while her mind's still fertile.

[Shankly998]

I know we have a few scientists on the forum so I wanted to know how they view their roles with the emergence of AI.

Is it possible or even feasible that scientific researchers for example could be replaced by robots?  I understand that maybe the data analysis could be but I'm not entirely sure.

My reasons for asking is my eldest granddaughter now 13, loves science especially bugs and insects which I'd like to encourage. 

From the research I've done into career options it seems that entomology is a very specialist field especially in the food and agriculture  industries but is it likely to be one that's taken over by AI due to the scarcity of scientists?

Absolutely possible our intelligence as humans is driven by a particular arrangement of atoms there's no law of physics that prevents this being reproduced in non-human form. It's just a question of when not if could be 5 years could be 50 years no one really knows.

[redbyrdz]

Not sure about needing a PhD. I have one, as do most of my colleagues, but it is not a job requirement, and technically you wouldn't even need a degree.

I think the academic route is definitively the most common, but for example one of my colleagues started as a lab technician, and then studied and did his PhD.

I wouldn't think about a PhD at this stage though! I'd see that she gets A-levels, because that's hard to do later. Then either go to uni, or see if she can get a technical job (looking after plants in a research lab is a technical job too), if she's more practically minded, and not that keen on studying. She can then always go to uni later, with some work experience already.

[redbyrdz]

The problem is the jobs that are almost entirely mundane and repetitive and the lack of sufficient gainful employment opportunities elsewhere.  Of course, we could all just reduce to 30 hour working weeks, then 20, then 10... as the spoils of humanity's innovation are spread evenly amongst us all 

I share your scepticism. There is this quote that someone postedthe other day, "A future where the robots paint and write poetry, while humans labour for minimum wage is not the future I want." I think the problem is that some of the mundane jobs are actually complex and involve multiple senses, and it is just cheaper to employ a couple of humans, than to build a highly advanced robot.

For example, there used to be automated car washes at many petrol stations, but they have all closed down, and you can get your car washed for the same price in a manual car wash.

Or another example (one for Debs) - how easy do you think would it be to build a robotic hairdresser? There are so many things that go into that - an experienced hairdresser will likely know with one touch of someones hair what shampoo it needs, and which styles it is likely to take. At the same time, they already have ideas what style would suit that person, taking both body shape and personality into account. All before themanual task of washing and cutting hair, which requires quite a bit of dexterity and fine touch. And in the end, many people actually come for a chat and social interaction, and the haircut is just secondary.

My fear is that AI and robotics will cause mankind to split further, a bit like in medival times. There will be those that enjoy the spoils of the machines, with meaningful jobs where they tell them what to do and design new ones. On the other side will be those for who little changes, they'll work long hours, and get drunk after, with no real progress. The problem is, it just costs very little to make a baby, and they grow by themselves just by feeding them, without needing scarce resources. And with over 8 billion people in the world, they are also easily replaced. A highly-advanced robot likely costs what a cheap human earns in about 20 years, so as long as our world is governent by capitalism and economic principles, mundane human labour will continue.

[TepidT2O]

AI is already found use in identifying new antibiotics.

It will also be really good at finding molecules that will fit the active sites of enzymes. I also know that they e used it to find more successful retrosynthetic pathways for making medicines

[reddebs]

Not sure about needing a PhD. I have one, as do most of my colleagues, but it is not a job requirement, and technically you wouldn't even need a degree.

I think the academic route is definitively the most common, but for example one of my colleagues started as a lab technician, and then studied and did his PhD.

I wouldn't think about a PhD at this stage though! I'd see that she gets A-levels, because that's hard to do later. Then either go to uni, or see if she can get a technical job (looking after plants in a research lab is a technical job too), if she's more practically minded, and not that keen on studying. She can then always go to uni later, with some work experience already.

That's really helpful info thanks.

I think for now it's more about letting her know she could get paid for messing about with bugs as I'm guessing it's not something that would come up in a careers lesson. 😁

[redbyrdz]

That's really helpful info thanks.

I think for now it's more about letting her know she could get paid for messing about with bugs as I'm guessing it's not something that would come up in a careers lesson. 😁

Haha, no, probably not!  I remember putting that I like being outdoors and that I'm concerned about the environment on my "career choice" quiz, and it suggested I should become a "waste removal operative", aka binman!

I think I'd encourage her to do as much science as she can in school, and (in a couple of years) see if she can do work experience somewhere that does research, if she is still interested then.

I'd think there are probably quite a few research fields that look into insects, from biology to agriculture, to healthcare, and also some materials engineering.




...in the meantime, get her to help you keep all the bugs and crawlies of your plants in the garden.

[reddebs]

Haha, no, probably not!  I remember putting that I like being outdoors and that I'm concerned about the environment on my "career choice" quiz, and it suggested I should become a "waste removal operative", aka binman!

I think I'd encourage her to do as much science as she can in school, and (in a couple of years) see if she can do work experience somewhere that does research, if she is still interested then.

I'd think there are probably quite a few research fields that look into insects, from biology to agriculture, to healthcare, and also some materials engineering.




...in the meantime, get her to help you keep all the bugs and crawlies of your plants in the garden.

We've had lots of bug hunts in the garden this week that's she's enjoyed and obviously we've discussed how important it is not to use pesticides, to encourage all pollinators into the garden and keep as much space as possible untouched and wild.

As much as she still enjoys it and is obviously interested it seems she's veering more towards the arts and becoming an interior designer.

I can maybe swing it with nobody will have the money for interior designers but the world will need scientists more than ever 👍

[PaulF]

Nobody I've spoken to thinks we can make a 'helmet' to do a gents cut.  I'd deffo use one . I'd imagine 'barbers' going to offices and just doing an entire office in an afternoon. And none of the patrons having to waste time getting their hair cut!
Somewhat different as you  say for the far more complex styles out there.

As a few of you have insight into these things. How hard do you think it would be to train a computer to spot when a dog squats in the garden to pee?  This destroys the grass and is easily fixed when a bit of water is chucked on the spot (apparently there's a 12 hour window).  Our dog makes a very distinct pose.  Similar to when pooing, though as we have to go sort that out too, a screen showing where she's done business that day would be ace.  Sadly I do think , it'd be cheaper (and definitely easier for now) to just detect movement in the garden and send it off to somewhere with cheap labour to identify!

[[new username under construction]]

[PaulF]

And that's before robots

[thaddeus]

And that's before robots

Robots from the future sent back to impose awful haircuts on any kids that might have been a threat to a future AI takeover.  It's arguably less humane than the Terminator approach!

That said, the mushroom cut of the first victim is enchanting.

[Party Phil]

As a few of you have insight into these things. How hard do you think it would be to train a computer to spot when a dog squats in the garden to pee?  This destroys the grass and is easily fixed when a bit of water is chucked on the spot (apparently there's a 12 hour window).  Our dog makes a very distinct pose.  Similar to when pooing, though as we have to go sort that out too, a screen showing where she's done business that day would be ace.  Sadly I do think , it'd be cheaper (and definitely easier for now) to just detect movement in the garden and send it off to somewhere with cheap labour to identify!

Ask this guy https://www.bbc.com/news/technology-67053171

[PaulF]

Ask this guy https://www.bbc.com/news/technology-67053171

Yeah, it's in my to do list .
I think ai / ml is going to be a leap nearly as big as the internet. Just so might boggling we can't envisage what it'll bring. Unfortunately I think it will be 'owned' by a few that will get unbelievably rich off it, rather than spread the benefits.

[duvva 💅]

Genuinely thought this was going to be a thread in which Al (Eeyore) argued against scientists and scientifically proven facts

A Very interesting thread, but I’m slightly disappointed

[Bullet500]

Generative AI (large language models like ChatGPT, image generators etc.) is most talked about these days. These are nothing but calculators for language or images. They can't replace actual scientists. That said, they can help simplify mundane tasks for scientists.

[PaulF]

Isn't a large proportion of what scientist do very mundane? I assume a lot of it is manual and observation in a lab. So not a lot AI can do there. Ai and robots though...