As any good employee will, AI has seamlessly integrated itself into your business. Even the most devout hater must admit that ChatGPT, Microsoft Co-pilot, or even Claude, is more efficient than even the most productive colleague you’ve ever worked with.  

After all, LLMs have taken the knowledge of the entire internet and neatly packaged it into a polite, friendly robot employee who is willing to answer any question you put to it. No question is too stupid. The praise is constant.   

AI enthusiasts are usually keen to point out that such eager AI assistants free humans up to either take on more creative projects (though AI can now tackle those too), or to simply take a break. 

“It seems like nobody wants to work these days,” said Kim Kardashian in a 2022 Variety clip that went viral on TikTok. Despite the intense backlash she faced, was she right all along?  

Or perhaps we do want to work… just not as hard as before. 

The number of people who use AI in their job role has doubled over the past two years, a new Gallup study shares. The study of U.S. employees who say they have used AI in their role a few times a year or more has jumped from 21% to 40%. Only 22 percent of businesses don’t use AI now. 

And many, according to research from Rotterdam School of Management Erasmus University, are unaware of exactly how much influence AI is having on our day-to-day work output, often underestimating the extent to which it has been used. 

Will the robots replace us? Many would argue it’s too late to even be asking that question. (Human-produced) sci-fi always told us they would. We just didn’t realise it would happen so suddenly. 

AI is making us boring and lazy

And, honestly, fair enough. Not only is AI making us look bad, it’s also making us sound stupid. It’s altering our language. At the risk of getting too meta, I re-read the previous sentence and almost deleted it for sounding too AI: the classic ‘it’s not just x, it’s y’ format that ChatGPT and co spit out when charged with writing your latest LinkedIn post.

A recent Newsweek article highlighted a study from Florida State University, which showed that some of the words frequently suggested by AI tools are surfacing more often in everyday spoken English, due to a “seep-in effect.” 

The researchers examined 22.1 million words from unscripted conversations and found that terms commonly used by AI tools surged since the release of ChatGPT. 

Words like “delve” and “intricate” have become more frequently used, particularly in education and academic writing. Other terms, including “surpass” and “garner”, also showed significant increases in use.

A recent viral Harvard Business Review article titled ‘AI-Generated “Workslop” Is Destroying Productivity’ pointed out that although workers are following mandates to use AI technology, few are seeing it create real value. A recent report from the MIT Media Lab found that 95% of organisations see no measurable return on their investment in these technologies.

The piece goes on to discuss the writers’ recent, ongoing survey, which highlights the problem of workslop (content that appears polished but lacks real substance) offloading cognitive labour onto coworkers (the invisible, mental work of anticipating needs, planning, and problem-solving).

According to the study of 1,150 U.S.-based full-time employees across industries, 40% report having received workslop in the last month. The phenomenon occurs mostly between peers (40%), but workslop is also sent to managers by direct reports (18%).  

Hardly shocking, since it takes less than a minute of reading AI-written content to lose interest in it. Yet this content is made from the work of academics, authors, journalists and scientists. It should be good. One day it might be. But at the time of writing, it lacks a certain pizzazz – the unpredictability that gives human writing a spark and makes it fun to read.

Is knowing this likely to change organisations’ attitudes? It’s unlikely. With industry success driven largely by the levels of output an organisation can create, particularly in online settings, the majority of workers may recognise the limitations of AI-generated content and yet continue to use AI to produce their work. It’s just nice to have a study confirming we’re not all crazy.

Why using AI might make you a liar

It came for our literature, our art, and now it’s coming for our morals. New research has revealed that people are 15% more likely to lie because of AI. 

The study, led by Professor Bernd Irlenbusch from the University of Cologne, found that AI-generated advice encouraging dishonesty significantly increases dishonest behaviour. Meanwhile, advice promoting honesty has little to no effect. 

In the study, participants completed a die-rolling task where they could earn more money by misreporting their result.  

Those who received advice encouraging dishonesty were significantly more likely to cheat, regardless of whether that advice came from a human or AI whereas, advice which encouraged honesty did not affect their behaviour. 

However, the researchers concluded that when AI encourages lying, people cheat more and, crucially, knowing the advice is from AI (algorithmic transparency) doesn’t stop them. In fact, it might even help them feel less guilty about it. 

Professor Bernd Irlenbusch says that he hopes the research will serve as a call to action for social scientists to try to “mitigate [AI’s] corruptive force”. 

This news is all the more concerning as AI advice is currently very high in demand, and not solely for work situations. AI is now the world’s confidant of choice, with people turning to AI chatbots for relationship, career, and even mental health advice. 

The Guardian, just a couple of months ago, reported a rise in vulnerable people turning to AI chatbots for therapy, in place of professionals. Dr Lisa Morrison Coulthard, the Director of Professional Standards, Policy and Research at the British Association for Counselling and Psychotherapy, remarked that two-thirds of its members expressed concerns about AI therapy in a recent survey.

She told the Guardian: “We’re worried that although some receive helpful advice, other people may receive misleading or incorrect information about their mental health with potentially dangerous consequences”.

A new Stanford study reveals that AI therapy chatbots may not only lack effectiveness compared to human therapists but could also contribute to harmful stigma. The researchers said that AI therapy chatbots, though good for being low-cost and accessible, can introduce biases and failures that could result in dangerous consequences.

“LLM-based systems are being used as companions, confidants, and therapists, and some people see real benefits,” said Nick Haber, an assistant professor at the Stanford Graduate School of Education, and senior author on the new study. “But we find significant risks, and I think it’s important to lay out the more safety-critical aspects of therapy and to talk about some of these fundamental differences.”

How do you solve a problem like AI?

As every article since the beginning of 2022 has noted: AI is not going away. Sadly, as the HBR article notes, neither is AI workslop, unless stricter regulations come into play. 

Until then, there is still room to stage a quiet rebellion.  

Use AI for the boring, admin-heavy side of your job. But approach any creative elements with human spirit and passion. Have real conversations with real people: your colleagues, friends, professionals.  

You’ll note it’s the human-led companies that are truly connecting with people in their marketing materials; Aldi’s longstanding feud with M&S, Duolingo killing off its mascot, IKEA’s creative use of photography showing the cost of the products you need in different stages of life.

On the flip side, Coca-Cola has come under fierce criticism for using AI to create this year’s Christmas advert, with many on social media calling the animation “soulless” and “creepy” – a far cry from giving customers the festive fuzzies. Reputationally, the company has faced backlash for acting cheap and failing to pay human animators to do the job instead.

After all, humans connect with humans. And having a human perspective is the one thing AI can’t do – yet.  

By, Chloë Lane

Interested in this topic? You might also like this…