We are being told by finance bros, Twitter cretins and LinkedIn lunatics that AI (a buzzword that can mean almost anything now) will displace or replace anywhere between 40-90% of workers, enhance our productivity and make us more efficient, that it will free us up from hard work and give us more leisure time, and even more absurdly, that it will create a world of abundance for all.

Let’s break those claims down.

It is impossible for “AI” or robots to do tasks that require the level of dexterity and flexibility that only the human mind and musculoskeletal system can pull off, even ChatGPT says it would be impossible. In real life, unlike science fiction, robots need a degree of bulk to be stable and are not good at self maintenance. AI, being software, will always be buggy and the more tasks you try to teach a system the more buggy and resource hungry it becomes. If robots and AI could displace significant numbers of workers it would come with reduced reliability, reduced dexterity and increased unpredictability in many fields.
Microsoft, Meta and Google are talking up a big game about AI, but just take a look at the state of their platforms and Windows 11’s bloated and buggy condition. Google and YouTube are happy to host fake and scam ads so their moderation tools are failing to detect wrongful activity, unless they are allowing it. Instagram is infested with bots and sex pests - their AI moderation doesn’t protect users. Windows 11 still randomly crashes, looks like it was designed by a 12 year old in Microsoft Paint and users are already trying to uninstall or remove the Bing Chat bloat. That’s an operating system under development for almost five decades and it is a mess, so it’s not hard to imagine how bloated and buggy their future Godputer would be like in practice. Keep it far away from military bases and defence departments.
The people making such claims don’t have domain expertise of all the jobs and sectors they are talking about. They are salesmen, newsletter shills and report writers who attach themselves to whatever the latest trend is. Some of them were raised by house servants or at the top of a caste system, so they never learned to respect working people anyway. If their reports and posts are super optimistic and buzzwordy, and if they fail to mention the technical limitations and implementation problems of any new technologies, it’s because they don’t really know what they’re talking about. They’re no different to Deepak Chopra talking about quantum physics.
Many sectors are already operating at peak efficiency. We know that because we produce far more goods than we need and generate tons of food, electronic and clothing waste. We actually need to produce less, produce on demand, have more local production, higher quality more expensive long lasting goods, more locally repairable goods, and more just-in-time production and shipping. That’s something the polluting Sheins of the world don’t want to hear about. AI does’t solve this problem. Human willpower and cooperation solves this problem. An AI can suggest people take action against over production, slavery and pollution (things we already know) but it requires actual people to make the decisions and do it.
Whenever sectors do use automation to increase efficiency, the time saved is filled up again by producing more goods, more content, more projects and expanding product lines. Employees do not end up working less, just differently. This is exactly what we have seen in creative workflows. As a production and post-production creative, I have used and implemented everything from Photoshop Actions to machine learning based tools in our workflows to speed up work and reduce mental stress. The result of efficiency gains allowed companies to ask us to produce more content. A decade ago we used to produce about 3 images per product. Today we are likely to produce up to 6 images per product and an optional video. Implementing machine learning and automation isn’t plain sailing either and often comes with bugs that are never fully resolved.
At the pharmacy where my brother works they recently installed a state of the art robot for stock tracking and dispensing drugs. It didn’t displace any workers and requires onsite and remote support whenever there’s a hardware or software issue. That’s just how robots are.
“AI will free up our time so we can create art.” Not everyone wants to create art, but remember that one going around on the socials? It didn’t age well considering the web is now being spammed by AI art that anyone can produce and often rips off the styles of well known artists. Generative art looks attractive at first sight because our visual cortex is experiencing something ‘new’ but within moments a kind of dire existential dread sinks in, similar to when you get a robocall or a bot sends you a DM. The images are bland, lifeless and have a ghoulish vibe to them. Apparently even AI systems prefer real art and real photography to the generative kind because when they are fed only generative art their abilities begin to decay.
“Generative AI democratises artistic creation” is another one we sometimes hear, and at first glance it appears to be a true statement, but democratisation of content creation already exists with the plethora of options available. With generative AI (especially in the cloud) you are getting centralisation, it enriches chip makers, rent seekers and energy companies. It pushes up the cost of living while pushing down the cost of labour. It encourages talentless soulless executives and shareholders to tell artists their skills are near worthless. It reduces the quality of creative production and increases the ease of which spam can now be generated. It contributes not only to the enshitification of the web but the whole human experience. LinkedIn is full of 40-something men who generate images of women of colour (aka they’re saying please don’t give work to real women of colour) and claim they are part of a youth movement democratising content creation. They tried the same trick with VR, the metaverse, NFTs and crypto, by claiming they were democratising finance and being inclusive. They were largely rejected by society and then pivoted to AI, after they had caused millions of people to lose money.
We could already have a world of abundance. It is the wealthiest who rig economies and create artificial scarcity in order to drive their wealth higher. They’re doing the same thing with AI by driving up the cost of using software or playing games, increasing energy consumption and making energy scarcer, increasing pollution, and hanging the threat of AI over the heads of workers to scare them into complacency and obedience.
Driving a car is something a teenager learns to do and for the rest of our lives we mostly drive sub-consciously because we rely on known routes, laws and landmarks to assist us. It’s when laws are ignored that bad things happen. The best full self driving offered still can’t consistently perform on the level of a law abiding driver. If AI can’t actually drive cars yet (and in some parts of the world it won’t be possible at all) despite 40+ years of development, then AI won’t be able to do all those jobs that require a lot more complex real time reasoning, fluid thinking and dynamic responses than driving requires.
If the consuming public were happy with bots replacing people, then athletes and sporting events would have been replaced by bots and virtual sports already. Why spend $20 million on a footballer when a team of computer controlled footballers can play sponsored advert filled virtual matches? Golf is an extremely wasteful and inefficient use of land and resources. Why can’t that end and be replaced with virtual golf? Big Blue beat Garry Kasparov at chess in 1997. Fast forward 27 years, there's still no audience to watch AI chess players play against each other in AI chess tournaments. The technology has existed for years, but consumers (fans) won’t pay for that. They will pay to watch real athletes struggle to win. Likewise, consumers will always pay more to read books written by real people, not chatbots. They want to build an emotional connection with the author, visit the author at a meet up, and get a signed copy of the book. A book is not just words on pages.
Finally, if the cost of producing something, whether it is art, literature or clothing, is closer and closer to nothing then there’s little incentive for customers to want to pay you good money for whatever you are offering. Your offerings are a McDonald’s Happy Meal at this point, or worse. The world’s economy can’t be made up full of Happy Meal and fast fashion equivalents. Every sector depends on diversification of goods and services, from high end and artisanal to the low end mass produced.

I end this blog entry with a video of a delightful lady who runs one of Tokyo’s many popular food joints ‘Onigiri Bongo’. Japan already has a few restaurants with robot staff (they are gimmicks), but a robot cannot make a thousand onigiri a day without health and safety hazards and causing a mess, cannot build a rapport with customers and cannot make customers wait in a line outside for an hour every day. Connections, traditions and craft are important.