Five things everyone should know about AI in the workplace

1. Generative AI is not a genius—it’s a predictor

When people talk about artificial intelligence, they usually mean machine learning. In machine learning, a model is trained on vast amounts of data, enabling it to identify patterns and make predictions or classifications. For instance, a model can be trained to detect suspicious changes in lung X-rays or to predict the weather.

ChatGPT and many other generative AI tools are based on a large language model that has been trained to use huge amounts of text to predict the next word or token. This allows it to generate new material for users.

Because generative AI relies on prediction, its ability to create truly novel content is limited. Rather, it generates new variations based on the data it has been trained on. For creative work, generative AI still needs human guidance to achieve meaningful results.

2. The impact of AI depends on the industry, role, and skills

Researchers have tried to identify the professions and tasks where the impact of generative AI is greatest. However, the phenomenon is so new that much of this remains speculative. 

Unlike previous waves of automation, what’s striking this time is that the entire wage spectrum is affected. According to researchers from Princeton, University of Pennsylvania, and New York University, high-paying professions may even be more vulnerable to the effects of AI. These changes are especially relevant for knowledge workers, who should take them seriously. The more a job involves processing information—and especially language—the more likely generative AI will impact it, while the jobs of, say, a bricklayer or a dancer will remain pretty much unchanged.

The oft-repeated mantra that “AI won’t take your job, but a colleague who knows how to use it might” doesn’t necessarily hold true. The biggest disruption is more likely to come through the creative destruction of the economy. AI enables entirely new business models, increasing competitive pressure on established companies, which may, in turn, be forced to cut staffing costs. Many established firms might find it challenging to fully harness the potential of AI.

3. Change is slower than expected—but it's something that society as a whole needs to prepare for

When ChatGPT was launched in 2022, it represented a significant leap—not so much in terms of technology but in user experience. The potential applications are easy to envision, and the tool itself is simple to use.

However, previous research has shown that major changes and benefits emerge only when new technology reshapes the entire operational logic of a company—a process that takes time. For example, when personal computers entered the workplace in the 1980s, their impact on productivity only became evident over time. Writing may have become more convenient—allowing corrections with the push of a button instead of using correction fluid—but true productivity gains come later, as organisations adapt and actually transform the way they work.

Listening to tech enthusiasts in the media, it might seem like the impact of AI on work is already immense. Yet in large companies, significant changes may take at least a decade to materialise.

One reason for this slow pace is the mentality of management. Executives often view new technologies through the lens of existing business models, focusing on how they can be improved. This limits the potential benefits of AI.

Another factor is the resistance of the organisation to change. Even if management comes up with new ways of doing things, it can be difficult to engage employees. Employees may understand that change is not positive for their own future and rationally put the brakes on change.

The realities of the labour market must also be taken into account. When envisioning new business models, it’s worth remembering that change will likely be implemented by the same people who were responsible for the old ways of doing things.

Despite the hype surrounding new technologies, broader societal discussion isbecoming increasingly important. Topics like education and labour policies, as well as the necessity of social safety nets, should already be part of the conversation.

4. AI does not automatically boost productivity

Economist Daron Acemoglu has estimated that if generative AI is used solely to enhance individual workers’ efficiency and reduce costs without fundamentally transforming a company’s or organisation’s mission and operations, overall productivity will increase by a mere 1% over the next decade.

From a competitive strategy perspective, using AI to compete on price is a risky path. It can easily lead to a "race to the bottom" phenomenon, where others also join the price war, ultimately failing to give anyone a sustainable competitive advantage.

One potential significant benefit of AI could be the improved availability of labour. For instance, we might be able to hire people who previously could not work in certain roles due to language barriers. This possibility was recently explored in a master’s thesis at Aalto University.

Generative AI also enables entirely new ways of working and better services. For example, if a doctor no longer needs to spend the majority of an appointment documenting medical records—because AI can transcribe and process notes from recordings—the doctor can dedicate more time to engaging with and listening to the patient. This kind of application is currently being studied in Finland’s Western Uusimaa Wellbeing Services County.

5. AI can make work more fun—but also more demanding

The benefits of generative AI for employees have been explored through surveys and experimental studies. These have shown that workers can complete some tasks up to a third faster thanks to AI—a massive improvement, especially for a technology that has only been in use for a short time! The use of AI assistants also appears to increase employee satisfaction.

In the long term, however, the cognitive demands of knowledge work may rise. It’s easy to assume that removing routine tasks simplifies work, but we often overlook the added pressure to do other, more demanding tasks. If your current workweek involves 30 hours of routine tasks and 10 hours of creative or high-level thinking, what happens when those proportions flip?