Will AI take your job — or make it better?
In this video short video published this afternoon, I compare AI in 2025 to the internet in 1995. Back then, just 5% used the internet. Now it's over 90%. Could AI follow the same path?
We explore what it means for work, power, and the way we run society. Change is coming — but will you master AI or be mastered by it?
This is the transcript:
Will AI come for you?
It's an open-ended question that I've just asked there, because what am I talking about?
I'm talking about two things.
First of all, will you use AI?
And secondly, will it come for your job?
Let's deal with both very quickly.
This chart is from the US Federal Reserve Bank, the St. Louis Fed, and what it shows is the growth in the use of the internet in the USA since it was created in the 1980s to pretty much today. And as you'll see, that growth was pretty spectacular.
I began using the internet in the early 1990s and was a director of IT companies promoting the internet dot.com boom by the mid-1990s. So, I was one of the early adopters, but back then, we were rare; around 5% of the population. And now usage is over 90% of the population.
AI is in that very early stage, just like the internet was in 1995, or maybe earlier.
But will it get to over 90%? I think it will.
And what does that mean for us? Because AI does automate tasks that previously we had to undertake, and we are already using it in the generation of these videos, by the way, for some of the mechanical processes behind them.
Will your job go as a result of AI? And if it does, what will you do, and how will we have to change the way that we organise society as a consequence if there aren't jobs for you to go to?
Now, people said this of the internet, by the way, and of course, it did not crash employment, it created new employment opportunities, and it's my suspicion that that is what will happen with AI as well, but we don't know what they are right now.
So, there's a big lesson to learn about AI, and that's to:
- First of all, learn about it
- Second, use it
- Third, master it
- Fourth, make it your tool and not your master, and
- Fifth, be prepared to change because the internet did change everything.
Can you remember, if you're old enough, when you couldn't Google something? The world was different from now, and AI is going to feel like that in 20 years' time.
So get ready. AI is coming for you.
Poll
Will AI change the world?
- Yes, we're all going to be exploited by big tech (43%, 109 Votes)
- Yes, it's going to wipe out vast numbers of jobs (18%, 46 Votes)
- Yes, it's going to change everything for good (16%, 40 Votes)
- I don't know (16%, 40 Votes)
- No, it's just another tech blip that we'll embrace with ease (8%, 20 Votes)
- I don't know and don't know and don't want to know (0%, 1 Votes)
Total Voters: 256
Loading ...Thanks for reading this post.
You can share this post on social media of your choice by clicking these icons:
There are links to this blog's glossary in the above post that explain technical terms used in it. Follow them for more explanations.
You can subscribe to this blog's daily email here.
And if you would like to support this blog you can, here:
Buy me a coffee!
If we had guaranteed basic income and perhaps also universal basic services then losing your job wouldn’t be a massive disaster.
It would be a bit like putting most of the population on tenure, which might be more interesting than everybody doing bullshit jobs they hate.
Lockdown showed us just how few workers are actually essential.
UBI means no public sector services, even if it didn’t start that way conservative governments would soon start cutting and privatising services using the argument that people should have the freedom of choice of using their UBI to buy services from the private sector. That’s certainly what Milton Friedman had in mind when he came up with the idea.
“if enacted as a substitute for the present rag bag of measures directed at the same end, the total administrative burden would surely be reduced.”
Friedman went on to list what he meant by “rag bag” : direct welfare payments, old age assistance, social security, aid to dependent children, public housing, veterans’ benefits, minimum-wage laws, and public health programs including hospitals and mental institutions.
That would have to be a very big UBI
A VERY big UBI.
Ain’t gonna happen.
No
It’s what proponents on the right argue for like Charles Murray (a somewhat unsavoury character) in his 2009 book ‘In Our Hands: A Plan to Replace the Welfare State’ which utilises a form of UBI and some dodgy calculations. More fanatical elements of the Tories will doubtless already be aware of the work of Murray and his ilk.
In the unlikely event a UK government went down the UBI route I wouldn’t expect the UK to start with that approach though, rather it would start small but conservatives with pound signs in front of their eyes couldn’t resist turning state functions over to the private sector and including them in things to paid for out of UBI, doubtless with increases in the UBI amount which wouldn’t actually cover the real cost for those who needed them most.
The Murray Book,https://books.google.co.uk/books/about/In_Our_Hands.html?id=6SW7AAAAIAAJ&redir_esc=y
Why let Friedman dominate your thoughts about UBI? There are other thinkers who present a more optimistic vision, such as Guy Standing. Of course, conservatives could make nasty cuts to an UBI down the line, but then they also do that to current social security arrangements and public services. People who value it, who I dare say would be almost everyone once an UBI was in place, would just have to step up to protect it, like they should do now. One big difference between what we have now and UBI is that the latter would be an entitlement as opposed to a ‘benefit’. This means it would be harder for evil Tories to take it away as its recipients could no longer be characterised as scroungers. Another reason, perhaps, that make it less likely to be introduced in the first place!
Good points Mathew and good video Richard.
Before Musk got into bed with Trump, I am pretty certain he went on record as saying UBI was the way forward. His actions while he was sidekick to Trump suggest otherwise.
Am pretty certain that Richard has spoken on a compass podcast saying UBI isnt feasible – yet – or at all.
Personally, I think based on everything going on in the UK, that there is going to be a jobs apocalyspe and by that point… Well… I dread to think what will happen. Everything may be getting run down for reasons we cant see yet.
Didnt Spain go with UBI during Covid.
I just cant see UBI happening because the press would be screaming about it and the vile culture we now have of the vilest opinions expressed by those who in power, look only at the world with cruel, pierced eyes would weaponise
it.
There was a time when I thought it may happen – we would get Ai to do all the jobs we dont want to do while we can live creative, productive lines. Now the Ai is going to take away that and its us that will be the slaves.
This is an issue I will get to
We had the equivalent of UBI in Canada during COVID. $500Cdn/week for those without work. It worked very well. The jobs apocalypse has already happened but in a disguised form of casual/precarious work, part time work, and wages below what it takes to sustain a person. Government subsidizes the corporate sector and small business through tax breaks and welfare/disability payments to those on low wages and low working hours. If we had a proper distribution of wealth, the biggest problem for the working class would be deciding what to do with all their leisure time (Keynes). We used to discuss this in our union in the 1970s. Workers are financially worse off now.
Let me give one area where it could change things: opticians.
I have worn glasses since the age of 8. The lenses (plastic) cost a bloody fortune and opticians and lens makers (& frame makers) seem to have a nice little number going. There is no reason why A.I. could not pass the exams to be an optician. I have identified equipment suppliers that make lenses & coatings – all automatically. Thus one could cover a country with a single A.I. running the eyes tests & what amounts to hole in the wall shops with a single technician for adjusting frames and fitting lenses. Could be/should be cheap as chips.
This is one example – there could be plenty of others.
Given it seems the pre-eye test by specsavers now seems to be unaltered by the optician this seems entirely plausible.
Although I voted don’t know, I’m not suggesting any ignorance on my part because I don’t really think that the consequences of this are predictable. I was also an early joiner of the internet, but in terms of access to information, at that time you could already have an encyclopedia on a cd, so it wasn’t a complete cliff edge.
I voted for “Yes, it’s going to change everything for good” based on my assumption that by “for good” you meant “irreversibly”. I didn’t mean to imply that I thought the results would all be beneficial.
That was the correct interpretation. Thanks.
I read too quickly and read for good as for the good. Probably would have changed my vote. Says quite a lot as I think more about it.
I voted that it will change everything for the good. But there are clearly risks and dangers that all of us, together with Government, will need to manage.
I’m currently using AI to help me make a case to my MP for a significant change in the way public finances are managed. So far I’ve got as far as securing her agreement that she will write to the Chancellor about it. I don’t think I could have achieved this without AI’s assistance. I’m not stopping there though, obviously, so I’ve used AI to help identify other steps I can take to press the case, and I’ve just sent another message. So my advice to anyone would be – embrace AI, but use your brain too. You’re going to need both.
Entirely correct.
Can we compare the internet and AI in anything but the natural progress of control, laziness and greed?
It’s a machine that will always be hungry .It will optimise consumption and increase environmental degradation!
AI already knows we are fucked!
But that’s another conversation.
‘AI already knows we are fucked!
But that’s another conversation.’
Hmmmm…… Sci Fi land. Asimov was thinking about this decades ago. He did invent the ‘Laws of Robotics” to give humanity a chance. But then he was writing fiction.
When AI can self replicate …..then we really are fucked. Redundant. Or at best servile.
AI only reflects what it has been trained on. I recently tried asking google for forthcoming dates that would have taken a few minutes to locate using various links, and it only gave me past dates.
For complex searches I now use perplexity.
I have used other tools occasionally, for specific purposes, such as Winston to check two texts for differences.
There are other possible specific uses in my business, but I am wary about letting AI make judgements.
I would never use it for judgements.
But it makes teleprompts from notes really well.
And I have trained it on 23,000 blog posts….
Thanks Richard, an interesting video.
I sometimes wonder whether the economy will be able to create enough new jobs to replace those being lost to advancing technologies.
In my view, education and especially re-education, will be crucial as people adapt to new industries and roles.
Hi Richard,
The second voting option “Yes, it’s going to change everything for good” has two entirely different meanings. Just saying!
Keep up the great work,
Guy
I realised afterwards….
There are an estimated 80,000 teaching staff in the further education sector in England.
Is artificial intelligence (AI) better? No.
But replacing the 80,000 would save £3-billion salary.
(Excluding pensions and the extra 120,000 support staff).
Is artificial intelligence good enough? Depends who gets a proportion of the savings.
I think AI is, or soon will be, good enough for many scenarios.
My teaching, not replaceable. Much of teaching, maybe.
Agree you first point Richard, but I am not sure how AI would get to grips with the hidden curriculum. Maybe teachers need not be fearful just yet.
Nuance.
That’s the Achilles Heel of it, so far.
I’ve been playing keepy uppy all my life. I was apprenticed as a compositor and was told I’d have a trade for life. I just finished my time in the early 80’s when the newspapers threw out all the typecases and hot metal machines – it was a little like getting up to dance just as the music stopped.
The first lesson I learned was the need for life-long learning and the importance of keeping abreast of all new technological developments. Of course I undertaken retraining to do the job using desk top publishing etc and new hobbies like digital photography necessitated keeping ahead of Apple, Adobe products etc. I’m lucky as I’ve never been afraid of educating and adapting, even to the point of undertaking a degree in adulthood. Many of my peers in the more traditional industries fell by the wayside; if their jobs weren’t replaced by technology then occupational hazards taken a toll on their health and most felt unable to adapt to the new technological world feeling they were ‘past it’.
So you’re right Richard – yet again we’re at the onset of yet another major technological paradigm shift and your advice is prudent irrespective of what kind of world we’re heading into – we all need to keep up as best we can. Ai is a bigger challenge, it has the capability of creating redundancies in much greater spheres of work. While it may be true that new opportunities will evolve, equally other opportunities will also materialise for those who have already devoured a greater slice of the world’s wealth, much of it made possible by the internet age. So globally the challenges that now face us all will be in our ability to make sure wealth and wellbeing comes to us all and the way that our politics is now shaping bears that out with the Trumps and the Musks of this world – this seems to be taking the form of them versus the rest of us. Concepts like UBI may be spoken about by some of them but there is certainly no movement by them to look at how equality for all might be achieved in a world of potential dwindling work opportunities – despite such lip-service being paid by them they’re certainly going in the opposite direction as Project 2025 illustrates.
Here, politically, recent developments with respect to Corbyn and Sultana organising a new party to oppose them show some glimpses of hope but it’s going to be a long struggle for us all I suspect. Such Ai induced changes are going to be dramatic and not only do we need to keep abreast individually but to learn to do so collectively and with some urgency given Trump’s ‘drill baby drill’ stupidity.
Thanks
My worry is that AI is going to be forced on us in our everyday lives before it’s been perfected to the level where it doesn’t make frequent and serious mistakes. But I also worry that it will be forced on us before the ramifications of using it—even if it works perfectly—are addressed.
Who will own and control AI? If it will be the same few faceless entities controlling everything for profit now, I am not optimistic.
If we want to embrace change with AI because it’s NEW, we also need to embrace new ways of providing for people’s needs. There is no point in rocketing ahead with AI, if we are going to remain stubbornly stuck in old ways of doing everything else. Or ignoring unemployed people who suddenly get cut off from basic things like food, water, heat, transportation, communication, health care, etc …because they no longer have an income, and can’t afford these things.
I want AI to empower me to do things I couldn’t do before. I don’t want it to make me feel helpless, stupid, discarded. I don’t want to feel terrified of what will happen if foolproof systems suddenly don’t work, and there is nobody out there to ask for help.
Accepted
With limited experience to date, I don’t fear AI. Yet.
But I do fear the datacentres…
https://steadystate.org/technocene-ground-zero-counties-face-off-with-data-centers/
Article by Amanda Jaycen for casse/Steady State Herald/July 2025
I urge everyone to read the blogs of Ed Zitron and his scathing assessment of the current AI bubble. It is currently more of an overhyped grift than anything else, and is no where near mature enough for us to know how impactful it will be. As for replacing our jobs, I doubt Sam Altman could name ten jobs, let alone know how AI could replace a person doing them, but that doesn’t stop him trying to sell his product. AI can’t take your job, but it’s makers can convince our bosses that it can. The other danger is the human centaur effect. In construction, we use AI to increase our productivity. There are too few people in the industry, so tools for efficiency are valuable. AI turns an engineer into a centaur, a human with horsepower. In other fields, this centaur effect is reversed by AI. Hospitals in the US sacking radiologists for example, and replacing them with an AI that makes so many mistakes that the radiologist is then forced to check the AI’s work at it’s pace. Instead of a human torso on a horse’s legs, it’s the other way around. Centaur effect.
I think this naive.
At the very least AI will tranform the jibs of millions of people.
In many cases, where people do routine – in coding, in law, accounting and some parts of medicine, for example – peoiple are goimng to see jobs disappearing very soon if they are not already.
To claim otherwise is just wrong. Sorry.
No apology needed. I should have emphasised the point that I think AI is being rolled out as an incomplete product and is being implemented incorrectly. The reverse centaur effect is a real phenomenon, it’s why Indian IT workers say AI stands for Absent Indian. We have too many cases of humans following up on AI work, not the other way around, as it should be. Until we break the AI hype bubble it is difficult to know how much of an impact it will actually have, though undoubtedly it will be huge.
Now we are very much nearer to agreement.
@ Tom B.,
I’ve read a fair bit of Ed Zitron’s output and I’d hesitate to recommend him to anyone. Not because I don’t think his points, re hype, aren’t valid; I think they are. Its that he rambles, can be repetitive and his articles are badly structured; he’s a very cumbersome read.
Gary Marcus, on the other hand, puts out concise pieces that get to the point:
https://garymarcus.substack.com/p/breaking-news-ai-coding-may-not-be
The dotcom boom was about selling goods to consumers. The social media revolution was about giving us tools for free and make us the product for them to sell to advertisers. The next step is to colonise the free web and make us paid subscribers to all the tools that were once took for granted as being free. And they know what price we will pay because they have our data.
I believe AI will change society in much the same way as all technological advances have altered things. What does worry me is that general education has not yet caught up with the Internet. How long before it catches up with the 20th century let alone the 21st.
Hi Richard
Very glad you made this post. I was thinking last weekend of emailing you to ask you to do so. I have a question on which I would like to hear your thoughts if you have time.
A little preamble if I may.
I have been a professional software engineer for nigh on 40 years. I have recently started to experiment with using AI in my workflow. I am by and large very impressed with it but some limitations are becoming very apparent. (I won’t bore you with the details) That said, in certain circumstances it improves my productivity massively. For example last Friday evening I had a coding task that would’ve taken me an hour or more and instead, in a couple of minutes, I created a prompt for the AI and the code it produced (in a matter of seconds) was perfect.
My question is more about political economy rather than technical capability (which will improve but may probably never supplant human beings.)
Consider a firm that employs 10 software engineers working a 40 hour week at £X per year. They produce bug fixes and new features at a particular rate. The software engineers don’t currently use AI assistance. The firm is profitable.
So the engineers introduce AI and productivity increases massively, they can produce the same amount of bug fixes and new features in one fifth of the time.
It seems to me that ALL that is being talked about is that AI is going to take our jobs i.e. the firm will sack 8 of the engineers thus getting the same amount of bug fixes and new features at the same rate. Same revenue but increased profits from reduced costs. I think this is the most likely scenario given the short termism nature of businesses, focus on shareholder value, perverse incentives (stock options) to management etc. and it is presented as this is the only possibility
What I am not seeing discussed are possible alternative scenarios. As described above the firm captures 100% of the productivity gains from the use of AI but rather than firing 8 engineers it could continue to employ 10 engineers paying them the same £X per year but reduce their working hours from 40 hours a week to 8 hours a week. In that scenario the workers capture all the gains from increased productivity. Another alternative is to reduce the hours of the engineers to 20 hours a week resulting in increased supply of bug fixes and new features which will hopefully lead to more revenue and increased profit so the productivity gains are shared between workers and owners/shareholders
Why is nobody discussing this (or have I just missed it!) AI could be of benefit to all if the productivity gains are shared by all.
In short it doesn’t need to be a jobs apocalypse, we could all share in the gains
Is there a way to ensure this? Maybe regulation to limit the numbers of hours worked in combination with levels of pay? Something else?
There are all sorts of issues with AI capability in its present state which I’ll try to go into at another time!
You are asking questions to which I need to give much thought – as that is very obviously necessary.
But I also hear Keynes saying work would not be needed by now, and note how wrong to date he has been.
So, will work be different? A marketing company I spoke to said, yes it will be. They used to write websites and content. Now they write AI prompts and create AI agents, using code. Adaptation always happens. I suspect that is most likely.
But there are very big questions in all this – and no one can be sure what will happen.
An interesting article yesterday at the Bulwark by Jonathan V Last may add to your thinking about AI – https://www.thebulwark.com/p/the-washington-post-is-dying-jeff-bezos-will-lewis
Noted
That AI generated approach will not be used here.
I will use AI, but not for ideas or opinion.
Its towards the bottom here
https://www.national-preservation.com/threads/have-a-good-laugh-at-really-silly-things.586905/page-454
A post by Jamessquared #9079 about a letter in Viz – where else? and Meta’s attempts to do something with it because it cant realise whats happening
On a more serious level; of course is the amount of electricity & water AI needs
[…] I was intrigued by the results of the two polls I ran in AI as a consequence of last night's video on that subject. […]
AI will first (and is) take all problem solving jobs. This starts with form filling, menial digital tasks and moves on to medical diagnoses etc. The next step is automation of that process to the machine that does the work. The opticians was a good example but most design and build activities can be done this way. Then comes enabling and decision making. This is where I’m not sure LLM can do it. It still needs guidance but I’m not sure whether it needs to become sentient in some way to really do this. Of course the dark side is that we let it make the decisions for us anyway. Finally we have general purpose robots that will implement the needs of its AI masters….. What will be left is carers, cases where human contact is not avoidable. These are the lowest paid in our society. The best bet is we all accept it is coming and make ourselves aware so that the decisions made at corporate and government level tend towards the benefits to society rather than profit and power but I’m not hopeful.
Having seen AI rolled out already in a local authority, I am unimpressed. People go around AI and try to talk to a real person in the services sector at least. Personally, if faced with an AI helper I go around it as trying to it get it to understand what I want is time consuming – and that is what the operator of the AI wants – to slow you down – demand management. And I fear – make you give up and go away.
This going around AI is part of systems thinkers call ‘failure demand’ as customers try to get their issues sorted out as they perceive the service to be unresponsive – I see a lot of this.
So Richard has got his AI aligned with his requirements, therefore it is learning what he wants; but what is the badly thought out or applied AI learning when people give up or get frustrated?
Another issue is given budgetary austerity, what sort of AI are we getting – top of the line or the cheapo version?
My view is that human is better in a customer service situation – I am not convinced that AI can cope with the complexity of people’s needs. And finally to me, I do not trust those who create AI because they can imbue it with their prejudices and ignorance.
But we will still get it. So you have AI that will be controlled by good people and bad people.
Much to agree with
I hate automated response services that never have the options I want and AI is not going to do much better than this in many cases.