Is AI Actually Good For Productivity? Here’s What The Studies Say

In just a few short years, AI productivity tools have become a fundamental part of our workplace culture with many companies embracing their use wholeheartedly. Meanwhile various AI studies have come forward showing a wild array of results with some reporting dramatic ROI (return on investment) while a handful suggest they may actually be slowing us down.

So what’s the real story? Are AI tools for productivity transforming the workplace for the better or is it all too good to be true?

For this deep dive, we’ll be digging into some of the most prominent studies on the topic. We’ll be examining their conclusions and assessing the real impact of using AI for productivity so that you can make an informed decision about the tools you use.

Artificial Intelligence and Productivity

On the face of it, it seems pretty clear that AI does contribute to improved productivity in many cases, but with some notable caveats.

For example, in a 2023 study conducted by the Harvard Business School, researchers found broadly positive results about AI’s ability to enhance tasks across a variety of skill and experience levels. In fact almost every study we’ve examined corroborated this evidence to some degree.

It’s pretty undeniable that there are many use cases in which AI tools for productivity are both effective and practical. That said, a lot of these studies also highlight serious downsides to overreliance on these tools.

One fairly common point of agreement between the various studies was that the greatest gains were made by people with a lower skill level. That’s not to say highly skilled individuals didn’t see benefits but there is a pretty apparent falloff.

More concerning results suggest that using AI as a part of your workflow will make it harder for you to function without AI while seriously harming your motivation.

Some AI Studies Raise Serious Concerns

A recent MIT study titled “Your Brain on ChatGPT” caused a lot of media buzz by suggesting that consistent AI users may “accumulate cognitive debt” through over-reliance on these tools. Put simply, the results suggest that as you offset more of your work to an AI tool, your mental muscles may atrophy.

Not only does this suggest that users might become dependent on these tools, but it also has the potential to stunt growth and learning, as the paper is quick to point out:

“Over four months, LLM users consistently underperformed at neural, linguistic, and behavioural levels. These results raise concerns about the long-term educational implications of LLM reliance and underscore the need for deeper inquiry into AI’s role in learning.”

And that’s not the only major study to come out this year with damning implications for AI productivity tools. In April this year, researchers from Zhejiang University in China released a paper suggesting that while generative AI does make people more productive, it also has a poor impact on their motivation.

According to their research it’s quite clear that while both the quality and the speed of work is dramatically improved by using AI, that comes at the cost of job satisfaction. Participants across the four studies they drew from reported higher rates of boredom and lower rates of intrinsic motivation.

Are These AI Studies Reliable?

Before rushing to conclusions, there are some important points we should consider. First and foremost, the above studies represent a relatively small pool of research on a topic that is sure to get a lot more attention over the next few decades. It’ll be a long time before we can really know how accurate their conclusions are to work culture as a whole.

Furthermore, as some online have pointed out, concerns about technology making people lazier and less motivated to work hard are essentially as old as technology itself.

On the other hand, if we are to take these studies at face value, there’s still a lot we can do as individuals to avoid the pitfalls they suggest. You can do this by limiting your use of AI to organisation and research, engaging in more active methods of learning, and being careful never to automate the fun out of your work.

Of course, that’s not to say all these concerns are just hypothetical. While this field research is young, it’s nonetheless very important. Its implications can – and in some cases will – predict the future. That would be particularly concerning in the case of the final study we’d like to talk about.

Can AI Have A Negative Impact on Productivity?

Up until now, all the studies we’ve discussed have generally agreed that the impact of AI does improve both the speed and quality of work being produced. Now we’d like to take a moment to talk about a study that has bucked that trend, why many people are worried about it, and what it could mean for regular AI users.

Research published earlier this year by METR found a very different story to other studies in this field. In a controlled trial of experienced software developers, they discovered tasks took 19% longer to be completed without AI than with it. Perhaps even more worryingly, the developers themselves seemed completely unaware of this.

The developers in the trial initially predicted that AI would speed them up by 24% and even after the trial they claimed it had made them 20% faster.

All in all, there’s a pretty clear conclusion to be drawn here. If this research is to be believed, AI isn’t just making us less productive, it’s tricking our brains into thinking we’re getting more done.

By now it should come as no surprise that this paper caught headlines. It’s a shocking story if true but, as with all the other research into this field, we have to take it with a grain of salt.

First of all, the sample size was small. Very small. Only sixteen developers participated in the trial which is a vanishingly small number compared to the number who currently use these tools on a daily basis. Furthermore, this research is specific to software development, while AI is used across all sorts of tasks.

This isn’t to say the study is completely meaningless. At very least it’s a valuable thought exercise and demonstrates that people are more than capable of biasing themselves when it comes to productivity.

At the same time, given that a study like this one simply cannot hope to address the breadth of uses these tools are being put to, your own experiences will be a far better indicator of what AI means for your productivity. The responsibility is on each of us to be honest with ourselves about how we’re using these tools, why we’re using them, and whether they’re really helping to get the job done.

Finding The Best AI Tools For Productivity

The effectiveness of an AI tool varies a lot depending on the use case. If you want to use AI in a way that’s really going to improve your productivity rather than limit it, then it’s important to identify the right use cases and tools.

As a rule, using AI for productivity is at its best when human input is a fundamental part of the process. For example, if you’re someone who struggles to find the right tone in emails, then an AI tool could speed this up a lot. All you need to do is give it your notes and read through the output to verify that its response matches up to your needs.

To put it simply, AI tends to be at its best when you use it to collaborate rather than letting it take the wheel.

It’s also really important to understand that different AI models are good at different tasks. If you want to find the best AI productivity tools, then you’ll probably have to conduct a fair bit of your own research.

Final Thoughts

As we’ve hopefully made clear by now, AI studies represent a very young field of research and one that isn’t without its controversy. Right now, almost every major tech company has their own AI model meaning there’s a lot of financial interest in the space. This only serves to deepen the problems that already exist with any research that might take place.

Some studies will inevitably be looking to find specific conclusions for financial reasons. Others will be performed without any biasing at all but will then be filtered through a media landscape that is driven by ‘clickable’ stories – the more extreme the headline, the better. All of this only adds to the challenge of finding practical, productive ways to use AI.

But let’s get back to that initial question. Is AI actually good for productivity?

The four research papers we’ve examined today represent only the tip of the iceberg when it comes to this field, but if we can draw any conclusions from them, then it’s that artificial intelligence is good for productivity in the right circumstances.

When all’s said and done, being an expert in your field is always going to be an advantage and that can only come with time and hard work. AI is very much capable of helping you become that expert so long as you know when/how to use it and are honest with yourself about the results you’re getting.