Artificial Intelligence – Will AI REALLY end ALL Jobs?
Or do we Just Need to Rethink Our Place in this Brave New World?
It was quite a sweeping statement from Elon Musk following the recent Bletchley Park AI Safety Summit when he commented to Rishi Sunak that “There will come a point where no job is needed.”
Now, much as I agree that there are certain jobs that AI is coming for, I have to say I’m not on the same wavelength as Elon in terms of the entire world’s workforce facing the chop and being overtaken by robots.
Not yet, anyway. At the time of writing, I am having my house painted. Watching the skill and dexterity of the decorator, it's hard to envision this task being done by an AI. Instead, I find myself imagining a wayward, wall-hugging, paint-spraying robot inadvertently glossing the cat. So, while I do not personally believe ALL jobs are at risk, I do believe that some most certainly are.
There are definitely certain jobs where AI is already starting to make a good run for employee of the month. And there are some that I (and seemingly countless others) believe the tech will eventually completely replace over the next few years.
So should certain professions be quaking in their boots? Will their AI imbued printer even bother printing out their P45 or will it just tell its sibling at HMRC directly? Or is it just a case of needing to look at this whole situation through a different set of specs?
Routine jobs
The fact is that a lot of jobs involve routine, time-consuming tasks. Tasks that take up valuable human time, which could potentially be repurposed to do something a whole lot more creative and helpful.
And it’s these routine, prescribed tasks that are already being replaced with AI. And in many cases AI is doing the jobs a whole lot better, and faster. And I wager it makes a lot less mistakes than humans too.
Take a tax investigator. Or a security patrol. Both do pretty much the same job… they scan for anomalies. One looks for irregularities in people’s accounts. The other looks for unusual behaviours in crowds. Both jobs could be done by a machine, a lot faster, and a lot more accurately. And this only has to be a good thing. HMRC uncovers more fraud, and crime is reduced. With money saved in the process.
And then there’s something very close to my heart: software development. I use a specialised version of OpenAI’s LLM in my daily software development role at Ban.gl. For the last few months, rather than writing code, I write a comment about what my code does, even before I have written it. Then, the AI takes over. I wait a second, and I receive a suggested code block back. The more I use it, the better it gets. It learns how I code and how the system I am working on is wired together. Sometimes this tool is producing a hundred hours’ worth of output in less than four hours’ worth of guidance by me. Seriously, I love it.
Plus, the confidence I gain from using this tool means I no longer shy away from implementing features that would stretch my ability and just take up too much time to research. Nothing now is off limits. All I need to bring to the table is imagination and the ability to verify that the suggested code is correct. My 30+ years of expertise as a coder is evolving into a role akin to that of a musical conductor. My code editing suite and AI tools are the orchestra. I just have to stand there, metaphorically waving my arms about, directing this symphony of technology.
I use my expertise to effectively guide this LLM. Pointing out when it gets things wrong and helping it refine its suggestions until they fit with my vision of what I wanted the software to do. Rinse and repeat.
Another aspect of software development that is often ignored or put on the 'To Do' list is documenting the code. It’s a bit boring, to be honest. At least I think so, and I rarely make a fuss about doing it. But now I just copy my code into ChatGPT, tell it what the code is for or what it represents, and ask it to produce a developer's handbook and an end-user handbook. It completes work in minutes that would otherwise take weeks to produce.
All that routine handed over to a machine. Fantastic. I’ve found more time to do less.
But do I think it will take over this part of my job completely? Er, yes, and frankly, good luck to it. I cannot wait. I don't believe I add much value in the writing of the actual code. There is a right way to do it and a bunch of other ways.
My value, as I see it, is my imagination, my creativity, my innate ability to imagine the rearrangement of pre-existing things to make a new thing. And of course my experience, knowing how things might work, what does work, what looks nice, what never works, what is needed, and what is often thought necessary but isn't. Working with the new AI tools, rather than without, is making me 100 times more effective.
It means I can let the machine do the boring bits, liberating my time to work on higher level stuff and be more creative and strategic, both in my own business and in my clients’, focusing on the bigger picture rather than minutiae. It could also free up my time to go for a walk, take a nap, or read poetry under the shade of a tree.
And this is where I think we need to rein in the ‘my job’s on the line’ panic a bit, and think more about what AI can do to make our work and personal lives easier and more fulfilling, rather than what we can do to try and avoid an AI management buyout. Because you know what, the genie is out on this and you’d be better off getting on board and buckling up.
But certain jobs ARE on the line, right?
Aren’t they always? Yes. I believe that anyone with a prescribed role is 100% going to find their job replaced by robots or LLMs at some point in the not too distant future. Within a decade I’d bet.
Accountants. Legal assistants. Pickers and packers. Call centre workers. Administrators. Receptionists. Tour guides. Support desk workers. Graphic designers. Non-specialist copywriters and editors. News reporters. And, like I said earlier, tax investigators, and security patrols. The list goes on.
But by ‘having their job replaced’, what I actually mean is, the routine, prescribed parts of it. There’s nothing to stop these traditional roles evolving with technology and blossoming into something a lot more fruitful. Or to stop a bright, savvy worker or entrepreneurial service provider from transforming what they do into something a lot more valuable.
It will begin as augmentation, as in the case with my day to day work, but progress inexorably towards a full takeover.
Let’s drill down in to a few more of these roles to see how they might be augmented with AI alternatives and what, if any, creative bits will be left to the fleshy professionals.
Take accountants as an example. Compiling year end accounts, submitting tax and VAT returns, issuing payslips, running reports… it’s all prescribed work that a machine could do without question.
But what about dishing out uniquely tailored advice that could save you money, make your business more profitable, help you make informed decisions and support you in achieving your growth ambitions?
I asked ChatGPT if it could give me personalised advice on my business accounts to help me be more profitable. This is what it said:
While I can provide some general information and tips related to accounting and finance, it's important to note that I'm not a qualified accountant and cannot provide specific financial or tax advice tailored to your unique business needs. For personalised guidance, it's typically advisable to consult with a human accountant who can analyse your financial situation, offer expert advice, and help you make informed decisions for your business. They can provide insights specific to your business's financial health and help you reach your financial goals effectively.
So, maybe the accountant lets AI do the boring stuff, allowing them to re-emerge as more of a qualified business adviser, with a good eye for figures. One that’s proactive and spends his or her time analysing each client’s situation and coming up with strategies to help them be successful.
Experts reckon that customer services the world over are already being overtaken by AI. IKEA, for example, reports that almost half of its customer calls are now handled by its AI, Billie. BT has also vowed to replace 10,000 jobs with robots.
For the first time recently, I noticed a novel use of AI in support triage. I emailed a company for support and immediately got a reply acknowledging my email but also suggesting links to articles that were related to the content of my support request.
I could click to say ‘yes, please cancel my support request as this is what I wanted to know’. It was slick, saved time and was undoubtedly the work of an LLM in the form of a digital assistant. It may have completely replaced a support professional, or it may be shouldering some of their burden, allowing them to focus on adding real value to those that really needed it.
Analysis by job site Adzuna has revealed that vacancies for graphic designers have dropped by 58% since 2022. There was no specific cause given, but with the rise of free tools such as DALL-E, one might have a smoking gun.
The image at the top of this essay has been created by DALL-E with the following prompt. ‘Please read this essay and make me a 1024 pixels wide x 554 pixels tall photo-realistic image conveying the essence of this article.'
I timed it… it took 17 seconds. It’s not as good as I would get from a human graphic designer, but it’s good enough for this use-case. I got what I wanted quickly, and it was free of charge. And for sure, I’m not the only one doing this.
And look what happened when ChatGPT met the Boston Dynamics dog. Good old Spot donned a top hat, moustache and googly eyes as it chatted with staff members in a British accent, taking them on a tour of the company’s facilities. So watch out tour guides, AI is coming for you too.
Professor Carl Benedikt Frey, future of-work director at the Oxford Martin School, Oxford University, said that technologies like ChatGPT are “very good at reproducing human interactions in the virtual world”. But that statement came with a caveat… “However, this is only going to make in-person interactions, which cannot be automated, more important,” he said.
So, maybe there’s a hint there that a spot of upskilling is needed. Because for sure, people in more junior roles are going to feel the AI pinch sooner than their more experienced counterparts. Where the tax man came for my generation’s Luncheon Vouchers, so artificial intelligence is coming for Gen Z’s Meal Deals.
The case for upskilling
Take copy writers. AI can certainly assist their roles by providing suggestions, checking grammar and spelling and generating content based on given prompts. But even ChatGPT itself says that there are several aspects where human copywriters excel.
Creativity for example. AI can generate content based on patterns and data, but does it really have the creativity needed to craft unique, engaging narratives? Can it evoke specific emotions or tailor content to resonate with certain audiences? This surely requires a certain understanding of human psychology that AI currently struggles to replicate.
And then there’s originality. Copywriters are responsible for creating original content, which can be a challenge for AI to accomplish without relying on existing data. But then again, it could be argued that copywriters rely on existing data. Data in the form of everything they have read which must be the seed to their original ideas. Unless it’s magic, of course.
ChatGPT had this to say on the subject:
While AI can automate certain aspects of copywriting and save time on routine tasks, the human touch remains crucial for producing high-quality, creative, and audience-specific content. Copywriting that goes beyond AI capabilities often involves elements of emotional connection, storytelling, creativity, and adaptability that are inherently human.
I’d go a bit further. Good copywriters, much like skilled doctors and accountants, put on their clients’ boots and walk in their world. Taking the time to understand their clients deeply. Exceptional copywriters learn to think like their clients. They can mimic their tone, use their preferred style of language, idioms, slang. They understand their clients' backgrounds and make instinctive decisions about what motivates their requests for copy. In this aspect, I don’t think we are quite there yet with LLMs.
Going back to ChatGPT’s choice of words, ‘Inherently human’, maybe it should probably say inherently higher skilled? Take, for example, a copywriter who doesn’t specialise in anything particular, who perhaps writes ‘prescribed’ keyword-led pieces designed primarily for search ranking. Self-styled SEO ‘experts’. They’re the ones that will be substituted by LLMs. But the copywriter who specialises in meaningful, niche, technical, emotive or thought-leading content will be a lot harder to replace.
Same goes for journalists. Reporters have already had the wind put up them courtesy of Google’s AI tool designed to write news articles. The tool takes information like current events, and automatically generates news content.
But your investigative journalist whose job it is to spend months or even years deeply investigating a specific topic and producing reports or writing books on it… OK they’ll get some handy AI help with their desk-based research and assistance assimilating their notes. But the essence of their findings will be based on their own initiative, empathy, insight and emotion. And those are skills that AI is yet to hone.
Those in pursuit of ‘job security’ could well need to start looking outside of traditionally ‘safe’ roles…
I suspect many choose their careers based on feelings of safety, of ‘job security’. Some will make their selection based on their own perceived importance of the title the job will give them. And others look for ‘A job for life’ or perhaps because it’s what their parents want. In my humble opinion, those strategies are now out-dated.
This article on job site Indeed, last updated March 2023, lists 31 careers with excellent job security. I have to say, I’m not sure they’ve hit the nail on the head with some of them, namely court clerks, tax examiners and librarians, which I think are all roles that could be done just as well by AI. Perhaps they may have been considered secure up until fairly recently. Until AI came along.
So what about solicitors and barristers, surely they’re safe? Even a blindly enthusiastic early adopter like me would be a fool of a client bringing nothing but a laptop and a subscription to ChatGPT to defend myself in court. I think that solicitors and barristers are, for the foreseeable future, safe. However, I am not so sure about anyone supporting them. A junior paralegal might need to start his or her own defence in the case of them versus AI.
AI in the form of legally trained LLMs is going to start getting some major air play in this domain. I’d imagine the aforementioned lawyers may well keep a close cadre of expert assistants, but who among them would not also utilise an LLM that has ingested every single case history since records began? An LLM that could recite every single line of law, without error, and that could instantly bring out precedent relevant to the facts of the case in hand?
AI is definitely shopping its CV around for this role. On that I rest my case.
And what about teachers? Surely no one wants to replace humans with machines in the classroom?
Now here is a profession that, in my opinion, should be legally required to start using LLMs as an aid in the classroom. LLMs excel at explaining almost any subject in any style and at any comprehension level you ask. I use ChatGPT for that purpose each and every day.
I usually start with, 'Please explain to me like I am an average thinking 10-year-old…' Then, I delve into the question that popped into my head while I was pulling up my socks: what exactly is a photon? What’s the difference between fission and fusion? Why did I think Rod Stewart was Scottish? What day is it today, and so on. It has yet to fail me. Using ChatGPT, I've acquired a vast, albeit superficial, corpus of esoteric knowledge in many areas that was previously way beyond my comprehension or reading ability.
Using LLMs in classrooms offers a great opportunity for kids to learn in a way that suits them. They can access information in a tone and language they understand, and now with tools like DALL-E, even in pictures. Which is a fantastic idea.
Another typical example of how AI could 10x or 100x the abilities of great teachers and their students all at the same time. And also an example of how the sidelining of not so great teachers may begin. If I could kick it off, I’d like it to start with all the teachers that told me, ‘Because I said so’, whenever I asked 'Why?' (And throw in the ones who made me drink sour milk, just for good measure.) And then move on to the ones who resist this technology..
I’ve already talked about accountants. Now there’s a traditionally ‘safe’ role that has forever been considered a top choice for lifelong job security.
But as I said, the prescribed parts of that role can easily be done by a machine. And if part of the would-be accountant’s career plan is to work for one of the big guns like PWC, Deloitte or KPMG for example, then they can bet their doubled entry bottom dollar that those prescribed tasks will ALREADY be handled by AI. And that these companies will be investing megabucks into AI adoption to streamline processes and improve profits. And they won’t give two abacus beads if it means laying humans off in the process.
I mentioned earlier that there was potential for an accountant to become more of a strategic business adviser, using their accountancy knowledge and mathematical skills to help their clients achieve their goals.
But not all accountants are going to have the skills needed to strategize in this way. They might just be great at working with figures. In which case, there are other roles… roles that are less likely to be overtaken by AI than they are fuelled by it. Like data analysis for example.
Now not only is that a career that has little likelihood of being usurped by AI, it’s actually one that could allow you to profit from AI, for example by specialising in artificial intelligence data science.
And that channels me nicely into my next chain of thought.
Can we finally break the glass ceiling of bias courtesy of jobs in technology?
History is littered with jobs prone to bias. Jobs that see invisible barriers preventing certain members of society from excelling in their careers.
Executive and leadership roles for example have traditionally been male-focused. STEM fields are the same, with women underrepresented in science, technology, engineering and maths careers, including our aforementioned accountancy. The same goes for construction. And then on the flip side, you’ve got teaching, nursing and caregiving which have for many years predominantly employed women.
The legal profession, corporate leadership and academia have been rife with racial bias. And some healthcare roles also face LGBTQ+ prejudice.
And this is why I think that if the future of work is really going to tick the equal opportunities box, then it has to lie in technology.
The fantastic thing about tech-led jobs is that there’s no glass ceiling. Anyone, regardless of gender, race, age, orientation, beliefs or location in the world can succeed in the field of technology. Talent and learning ability permitting, of course.
Unlike with some traditional jobs, there’s little to no bias. And now, if you can learn to leverage AI and LLMs in your role, and train others to do the same, you will be in such demand that your skillset will be the only thing that matters. How’s that for equal opportunity.
There are new jobs opening up all the time that could be a lot more lucrative and less biased than certain traditional roles, as well as being considered safe. It’s just a case of looking beyond the traditional, and seeing what’s out there. Upskilling or reskilling if necessary to suit the jobs of the future, rather than trying to hang on kicking and screaming to the jobs of the past.
Everyone, as members of the human species, has the same potential to be skilled in any area, because abilities aren't handed out unevenly among people.
Time for a career rethink?
It’s time, I feel, for certain workers to rethink their roles. So rather than slip into panic mode as ‘AI the grim reaper of jobs’ comes for your career, try considering how AI can be more of a joy bringer, sharing appealing benefits like higher wages for skilled roles, and more time to do less.
A few paragraphs back I touched on something that Professor Carl Benedikt Frey (Oxford Uni) said. Here it is again for memory’s sake:
Technologies like ChatGPT are very good at reproducing human interactions in the virtual world. “However, this is only going to make in-person interactions, which cannot be automated, more important.
In-person interactions definitely are important in certain situations.
How great will it be for teachers, doctors, nurses, therapists and accountants to get back to good old-fashioned one-on-one engagement, made possible with the help of a tireless companion in the shape of a well-trained, focused, LLM?
Take the role of the doctor. People of a certain age, myself included, will recall the concept of the ‘family GP’. These doctors knew us inside out. They knew our parents. They were completely up to speed with our medical histories (as well as our sporting interests, hobbies, education and work). We felt comfortable talking to them because they were like a knowledgeable friend.
Fast forward to the 21st century, and few will disagree that medicine has lost its personal touch. GP rapport has become a thing of the past. They have set times for each appointment, and they don’t seem to factor in a few minutes for rapport building and catching up with family news.
But what if they could hand some of that appointment slot time over to AI? Like the bits where they have to read your notes beforehand, and write new notes up afterwards. The bits where referral letters are issued.
So in other words, AI takes over the admin, freeing the doctor up to be able to spend more time with their patient. And AI could potentially do a better job – certainly a faster job and, dare I say it, a more accurate job - of analysing medical history notes and matching them with symptoms and similar cases to come up with a diagnosis and treatment plan.
Recently, I was copied into a referral letter from the Ear, Nose, and Throat Consultant to my GP, informing him of my, and I quote, 'sweaty spells'. What he meant to write was ‘dizzy spells', of course. An AI would know that labyrinthitis does not typically induce sweaty spells.
We all know that AI is making huge waves in the world of medical science. I for one would be happy to have an AI robot carry out surgery on my brain. I'm willing to grant an LLM access to my full medical history and DNA. This could help it predict illnesses I'm prone to, based on my unique genetic makeup, and uncover hidden causes behind any health issues I might face.
The same goes for air travel. For sure I’d be more than happy for an AI to fly me to my holibobs. If a computer can land a spent rocket engine, upright, on a boat, bobbing along in the middle of the beautiful briny sea, in an area just 91 metres by 52 metres, then it can land a big fat A380, with 22 wheels, and wings, horizontally, on a 60 metre wide, 3.5 kilometre long strip of flat tarmac. In fact, in adverse conditions, I’d prefer it.
Going back over the claim that all jobs are going to be taken over by AI.
My invite to the UK Summit on Artificial Intelligence obviously got lost in the post. As Elon won’t take my calls, I can’t be sure exactly what he meant, but I don’t think ALL jobs are going to be lost. Jobs will change and we will all have to change along with them. This has always been the case throughout human history.
I do not believe for one minute that humans will be surplus to requirements across the entire job market throughout the entire planet. That’s ridiculous. What I believe is that there will be a number of outcomes. Industries targeted by AI and experts who find a way to leverage AI in their roles will flourish. Those in roles that AI can be applied to who do not make use of the tech will be rendered obsolete. And those where AI is not applied will see little change. Think expert painter and decorator.
The ones who seriously know their trade and learn how to use AI to their advantage will see their natural talents multiplied. Conversely, those lacking niche, expert skills (or the drive to acquire them), and who avoid embracing emerging technologies, risk becoming obsolete in an AI dominated landscape.
So, how’s the future of work REALLY looking? Is AI set to replace the entire global workforce, or can we keep that Christmas turkey on order again this year?
No. And I’m quite sure about this. AI isn’t going to be singlehandedly responsible for the destruction of the human workforce. But if we’re not careful, we may be responsible for the destruction of our own careers. And this is why we need to take action now to prepare ourselves to work with this new technology. Everyone needs to consider how an AI might be able to do what they do, and also what they could do that AI cannot… yet.
I firmly believe that for every job that AI takes, more will be created. And those new jobs will be way more fulfilling than the ones they left behind. It’s just a case of finding our place in this brave new world.
If we’re savvy enough, we’ll use AI to augment our roles rather than sit back letting the tech steal them from under us. If we equip ourselves with the right skills, find a niche and become an expert in our field, we will remain on top.
The combine harvester – one of history’s biggest revolutions - is a good reference here. It reduced harvest crews significantly, resulting in huge savings for cash-strapped farmers. And whilst it saw unskilled jobs reduced, it created more skilled jobs that were also more lucrative, and made every agricultural worker vastly more productive.
And what about 3D printing? Rather than having forced a decline in manufacturing industry jobs, it’s actually led to new job opportunities in specialist areas like product design and 3D printing technology development.
Innovation is a natural part of history. It’s seen us wipe out devastating illnesses and improve treatments for others. Discover new planets. Invent tools that significantly impacted human civilisation.
And AI is just another one of those innovations. It just so happens it’s a pretty spectacular one. Perhaps the biggest of them all so far.
A final addendum… what exactly is the ‘meaning of life’?
Following the Bletchley Park Summit, speaking to Rishi Sunak, Elon Musk articulated that “One of the challenges in the future will be how do we find meaning in life."
Now I’m assuming by this he meant that if we all have our jobs taken away, there won’t be anything left to get up for in the morning. But is it really in our jobs that we find the meaning of life? Speaking for myself, definitely not for me.
Aren’t our jobs simply the vehicle that allows us the opportunity to find meaning in life? To travel, to explore, to learn, to play, to garden, to fish, to read, to create, to visit friends and family and to love?
And isn’t AI going to free up the time to allow us all to do more of this stuff? Absolutely it is. So let’s stop running away from it, and embrace it instead.
AI has the power to tear down glass ceilings and give EVERYONE the chance to enjoy their day job so much more. It has the ability to give people meaningful jobs that evolve with the times and play to their strengths.
But for those setting out on the path to securing a ‘job for life’, accountants, legal professionals, bankers and financial advisers, they’re going to need to seriously consider whether the career they’re choosing is suitably future proof. Will they be able to build expertise in that area quicker than the AI tools already operating within it? No, is the short answer to that in my opinion.
For those of us partway through our careers, we’re lucky. We have a head start. For the really old gits like myself with 30+ years in the field, we’re the luckiest. We will probably be useful for a while yet.
Of course, all the above is just my humble opinion. I’m no expert and all the bad ideas are mine and any good ones I probably borrowed unwittingly from someone else. Stephen Fry puts it’s so much better than I ever could and if you are still with me at this point I’d recommend watching his talk on AI here: https://futurist.com/2023/09/29/stephen-fry-on-how-to-use-ai-as-a-force-for-good/
One final thought. I have developed a tool which can help you get a better insight as to whether your role or career might be under imminent threat of AI takeover. You can try it for yourself, free of course, at https://ban.gl/oh-bother
I hope I have given you some food for thought.
Best wishes,
Anthony