The AI Revolution Isn't Happening
And we can prove it.
AI promises to revolutionise the workspace. By now, there is an AI tool to replace or augment every job under the Sun and send productivity through the roof. There are even entire industries, such as copywriting, that many predict will be entirely replaced by AI tools within the next few years. But what happens when you look past the hype and at how AI is actually affecting the work environment? Well, A study from the Upwork Research Institute has done just that and uncovered some worrying results.
This study surveyed 2,500 global C-suite executives, full-time employees and freelancers regarding their opinions on AI. What they found was a fascinating split in opinion and experience.
Overall, 96% of executive-level managers expected AI to boost worker productivity and reported high hopes for the technology. In fact, 81% of executive-level managers said they had increased demands on their workers in the past year, aligning with their roll-out of AI.
But the opposite was true for employees. 77% reported that AI had increased their workload. Why? Well, 39% reported having to spend more time reviewing or moderating AI-generated content. Furthermore, 47% of employees using AI said they have no idea how to achieve the productivity gains their employers expect, and 40% feel their company is asking too much of them when it comes to AI. As such, 71% of employees reported being burnt out, and 65% reported not being able to meet their employer’s demands on their productivity.
In other words, those pushing AI into the workspace expect it to make substantial productivity gains, but those actually using these AI tools are struggling to find such gains and, as such, are being overworked to make up for the deficit.
Now, this study was funded by a freelancer marketplace, so its findings might definitely be skewed, as AI directly competes with their product. But, there are third parties who have found similar results. For example, ING found that productivity will only increase by 0.1% as AI is embraced by the economy. Even MIT has published research highlighting how AI isn’t increasing productivity.
But we have been here before. Back in the 70s and 80s, the US rapidly adopted IT technology, which should have dramatically increased productivity. However, productivity actually dropped. We still don’t fully understand why this happened, so it is called the “productivity paradox.” Now, researchers are calling the misalignment of AI productivity expectations and alignment the “AI productivity paradox.”
But we actually know what is causing the misalignment this time. You see, AI is far less accurate and useful than many, particularly unengaged executive managers, think it is. In fact, we have some brilliant anecdotal evidence for that.
Take Amazon’s “just walk out” grocery stores. The idea was that facial-recognition cameras, shelf sensors, and AI would track what items a customer had taken, then charge their Amazon account once they left, negating any need for a cashier or self-check-out. This innovation was hailed as one of the first cases of AI directly replacing human workers and a way to lower the cost of operating a store. But, in reality, it really wasn’t. A recent report found that over a thousand remote workers had to be hired to monitor the video feeds and verify 70% of customer purchases, as the AI consistently got it wrong. This amount of labour isn’t cheap, even if it is cheaply outsourced overseas, and Amazon’s “just walk out” AI became significantly more expensive than simply hiring regular cashier staff. As such, Amazon has struggled to sell the system to third parties and has had to switch its own grocery stores to a fancy non-AI self-scan system instead.
Even more constrained examples, such as AI-generated code, have a similar performance issue. One of the internet’s favourite game developers, Jason Thor Hall of Pirates Software fame, brilliantly described this issue in a recent short, saying, “We have talked to people who’re using AI-generated code, and they are like, hey, it would take me about an hour to produce this code and like 15 minutes to debug. And then they are like, oh, the AI could produce it in like 1 minute, and then it would take me like 3 hours to debug it.” Again, the AI’s poor performance makes productivity gains while keeping quality high impossible.
But, surely, AI is going to get far better in the coming years, resolving this issue?
Well, probably not. As I wrote in a previous article, AI is starting to hit a point of diminishing returns. To continue the pace of development, AI has experienced over the past few years would require the amount of data AI is trained on and the amount of energy it takes to train AI to grow exponentially each year. Unless a major leap in AI training efficiency is found, AI development is likely to stall to a crawl over the next decade.
So, no, AI can’t revolutionise the workspace by skyrocketing productivity. The technology is too unreliable and requires significant human oversight for many use cases. Moreover, AI will likely stay this way for the foreseeable future.
Thanks for reading! Content like this doesn’t happen without your support. So, if you want to see more like this, don’t forget to Subscribe and help get the word out by hitting the share button below.
Sources: Upwork, Will Lockett, Will Lockett, ACM DL, ING, MIT, Brugel, Will Lockett