If you’re not confused in the world of AI, you’re not trying hard enough.
An old tech writer told me that artificial intelligence and machine learning have been around for decades. People today say that ChatGPT has given new life to AI. Still others are raising concerns that the breakneck pace of AI development will put many of us out of work. One global CEO even claimed that all promotions would be tied to AI knowledge. But if AI does all the work, who will we promote?
But last week, our confusion was compounded by the release of new research that shows that AI isn’t actually making employees’ jobs easier, but making them work harder.
Yes, that’s right. ActivTrak’s new study of the digital activity work of more than 164 million employees says, “Workdays are shrinking, but workloads aren’t. A study of digital activity in the 180 days before and after they started using AI suggests that AI has ‘enhanced’ employees’ work in nearly every category.
Time spent on email, messaging, and chat apps has more than doubled, and business software usage has soared by a whopping 94%. And ironically, this spike came at the expense of staff time spent on focused, uninterrupted work, with a 9% decrease for AI users and no change for the rest of users.
However, the study found that there is a sweet spot when it comes to using AI. Employees who spent 7-10% of their total working time on AI showed the highest productivity. However, only 3% of all AI users fell within this range. “It’s not that AI doesn’t create efficiencies, it just means that the capacity it frees up is quickly repurposed for other tasks,” said Gabriela Mauch, chief customer officer at ActivTrack.
Incidentally, this isn’t the only study questioning the impact of AI on workplace habits. Another study published by Harvard Business Review also found that AI is making jobs more intense, not less. Researchers Aruna Ranganathan and Shin-Chi Maggie Yeh talked about “workload creep.” This means taking on more tasks than the staff can sustain on an ongoing basis.
They say this can create a vicious cycle that leads to fatigue, burnout, and poor work quality. In short, AI tools have created a vicious cycle. “Certain tasks have become faster, expectations for speed have increased, and speed has increased workers’ reliance on AI. This increased reliance has increased the scope of what workers can try, and increased scope has further expanded the volume and density of work.”
The question then becomes how do companies plan to expand the use of AI within their companies? It starts with department heads driving the agenda and opening up data silos that have existed within the company for years. But the question arose when the CEO perked up and told all staff that promotions were related to their ability to use AI.
In a recent episode of the Rapid Response podcast, Accenture CEO Julie Sweet said that AI proficiency is an essential part of working and advancing at the company. “If you want to get promoted, you have to do what we do to run Accenture,” Sweet said. What we also know is that the company spent a whopping $865 billion last September on a six-month business optimization program that included employee reskilling.
Sweet was very tough on her approach, where the company gave staff a three-year gestation period to update themselves with AI tools. You can’t blame her, as Accenture has launched a three-year, $3 billion push starting in 2023 to integrate an AI-first approach. One of their goals was to double their AI talent pool to 80,000 professionals through hiring, acquisition, and training.
Does this mean all 770,000 of Accenture’s employees will have to be AI proficient in the new order? And if so, what about the job surplus that the use of agent-based AI solutions could bring? Perhaps that’s how Julie Sweet sees things. According to Gallup research, in the fourth quarter of 2025, only 38% of companies reported integrating AI to improve workplace productivity.
Sweet sees this as a continuation of changes that occurred decades ago. She believes that integrating AI into the workplace is a natural outgrowth of how computers became word processors in the workplace. Typing classes were replaced with computer skills, but are now being reskilled to include AI knowledge.
This is where the question of the AI employment apocalypse comes into play. Over the past few years, markets have been predicting major disruptions in the job market, and recent headlines included reports that several tech giants were preparing to make fairly large layoffs. However, a recent report published by Anthropic suggests that the worst may still be some time away.
The study, published a week ago in the journal Anthropic, suggests it may take a little longer for alarm bells to ring for some reason. Although “AI is far from a theoretical capability” and “practical coverage remains a fraction of what is possible,” some jobs may be at greater risk of “AI-driven replacement” than previously thought.
Jobs that could be at risk include computer programmers, customer service representatives, data entry operators, medical records specialists, and market research analysts, which are “most exposed” to AI capabilities, the post said. Once AI fulfills its destiny, investment analysts, software quality assurance (QA), information security analysts, and more could also face the axe.
Another factor revealed by the study is that even within these professions, the impact is likely to affect staff in the sector who are “older, female, better educated and better paid.” People at significantly lower risk include occupations that require physical presence, such as cooks, mechanics, bartenders, and clerks. Although the Anthropics called these “noble professions,” they did not pay particularly well. Perhaps there is a story for all of us that AI is slowly unraveling.
Perhaps there is a story for all of us that AI is slowly unraveling. Maybe there are multiple stories. Anthropic itself says that given all the noise about job losses, it’s hard to measure how many people have actually been lost to AI or other reasons. One example of this is how many companies have used AI as a reason to cut over-hired staff in the aftermath of the coronavirus.
Amid all these counterintuitive claims made by trigger-happy experts and so-called business thought leaders, one thing stands out strongly. Higher education and university degrees will need to be reshaped to the extent that all students will be forced to learn the ropes of AI tools while gaining knowledge in their chosen subject matter.
A very difficult question arises here. How many academics are actually proficient in using AI tools? Come to think of it, how many journalists are smartly and passionately leveraging AI to save time and produce better results?
