Dealing with artificial intelligence in the workplace

“Many people use Chat GPT to write texts. This makes sense, as it was the reason why it often appeared in the media. But far too few people use the technology to its full potential. There are many other opportunities to increase your efficiency or stimulate your creativity using generative artificial intelligence (AI).” This belief was the starting point of GPT Academy, established by Michael Bauwens from UCLL University of Applied Sciences, which aims to help SMEs deploy GenAI in a more effective manner.

How can AI improve the productivity and efficiency of knowledge workers?

Bauwens: “At first, AI could mainly be used to automate basic repetitive tasks. With the rise of generative AI, we have noticed that knowledge tasks can also be partially automated. On the one hand, this is positive news, as it allows us to perform difficult tasks more efficiently. But, on the other hand, it’s also a little terrifying. Because these knowledge tasks were a kind of ‘last stronghold’ of what only humans could do. Fortunately, by using this generative AI a lot, we have noticed that AI won’t simply replace us.”

What are some common misconceptions about generative AI?

Bauwens: “Many people see this technology too much as something human. The irony is that generative AI needs to be used in a very human way to get the best results. We need to talk to it as if we were talking to another person. However, large Language Models, like ChatGPT, are actually just very powerful word predictors. Another misconception is that companies such as OpenAI are inventing revolutionary technologies that didn’t exist before. That isn’t true. For example: OpenAI sometimes launches a new feature that is based on technology that has been around for years, but wasn’t released in such a way before. For a long time now, we have been able to search the internet with a browser without having to type literal keywords of what you’re looking for. It’s very easy, for example, to find articles about bicycles on Google without literally having to type the word ‘bicycle’. Keywords such as ‘traffic’ or ‘wheels’ also lead to articles about ‘bicycles’, because the topics are similar. Now that ChatGPT links these same search data to your own data, people are suddenly surprised. Even if it is primarily the same technology, just used differently. We have been using AI for many years in all our tools, hardware and software. But we aren’t always aware of it. And we don’t have to be. AI adoption is more effective if people don’t consider an AI application as AI.”

What are the consequences of generative AI for companies and employees?

Bauwens: “We have noticed an interesting paradox in this regard. We are very happy that these tools make our jobs easier and increase our productivity, but we also ask ourselves whether we will still be paid the same wage if we can perform half our job semi-automatically. For example, software company AFAS recently introduced a four-day week with pay retention, based on the belief that AI helps achieve the output of five days with four days of human work. But you can go even more extreme: what if we don’t have to work anymore because everything is automated? I don’t think we will be completely replaced by AI any time soon. But people will perform their job in a different way. And companies will have to take into account certain guidelines. In our study on responsible use of generative AI in HR processes, our starting point is that we want to take a look at how certain matters can be automated to free up more time for the real essence of HR. We want to make more time for the ‘human’ in ‘human resources’. We call this humAIn resources. Companies can interact with their employees in a more human and personal way by deploying the technology correctly.”

 

“AI really won’t simply replace us”

 

What are the important challenges?

Bauwens: “We have noticed that many people are concerned about privacy. An application such as ChatGPT sends a lot of data to American servers, which is far from ideal because of the GDPR and the EU AI Act. Companies wonder how they can use the technology without having to disclose their data to American companies. This is possible by working with local companies and using open source applications. But it requires some looking around. For many applications, Microsoft and Google are perfectly fine to use, but if privacy is really essential, you’d better be careful.”

“The second challenge is choosing the right tools. There are so many tools, start-ups and so much software launching new, smart features based on generative AI! It is impossible for us to test all of these new tools. That’s why we have adopted the rule to try to continue working with the rather popular tools, unless a very specific tool offers a very specific solution to an issue.”

How do you stay informed about the latest developments and the opportunities for your job?

Bauwens: “The latest developments aren’t always communicated in a way that is understandable for end-users. Developers are often the first to be informed about new trends or applications. Of course, there are always companies or people informing the end-users, but you need to be able to separate the reliable sources from the trendy influencers. I am more likely to believe the information provided by employees of a tech company than AI consultancy start-ups with no proven track record. And be wary of pushy campaigns aiming to sell you prompts.”

“At GPT Academy, we often present a GenAI Update with news from the last two to three months. And obviously you can trust us, as we are supported by a knowledge institution and have a track record and credibility to show for it.” (laughs)

What advice do you have for knowledge workers who want to address this with their employer?

Bauwens: “Don’t hide your use of generative AI from your employer. If you’re not honest about your use of the technology, your employer cannot create a good policy either. Your employer needs to know what is going on and where efficiency gains can be made. Otherwise, there will be blind spots in the policy and employees will not be supported. You might be less aware of certain privacy risks or you might be sharing various confidential documents in ChatGPT without realising it. If your employer doesn’t know you’re doing that, he can’t support and protect you on a policy level either.”

Author: Jan Deceunynck | Image: Shutterstock