The Impact of Artificial Intelligence on the Future of Work

Inspiration for this Newsletter: 
the "AI & the Future of Work" Meetup of March 24th:         

On the 24th of March, 2023, the Future of Work community organised a meetup on the impact of AI, in particular generative AI tools like ChatGPT, on the future of work, in particular the future of knowledge work. For this joint venture, we worked together with Elke Wambacq (managing partner at Dinobusters ), specialised in change management and human resources.

The meetup was in Dutch, but Elke and I created a sketchbook (well, she also used 7 AI bots to generate extra text and artwork) on how our joint venture on generative AI came about. You can download this sketchbook (in English) here ...

She also wrote a short summary (in Dutch) of the meetup, which you can read on the Dinobusters.be website .

The impact of AI on work        

It is clear that AI will have a huge impact on the way we work in the future, how we will work together with software robots, and how we can avoid an identity crisis.

A few years ago, we experienced a surge in interest surrounding Robotic Process Automation (RPA) tools. These tools were designed to automate repetitive human-computer interactions, such as copying and pasting between applications or screens, populating web forms, and performing basic calculations. Companies like UiPath, Automation Anywhere, and Blue Prism positioned their solutions as a means to execute mundane tasks more efficiently and cost-effectively. They also claimed that employees would be able to focus on more significant, gratifying and meaningful responsibilities, such as enhancing customer interactions and care. After all, when taking work away from people, it's essential to provide a positive outlook as well.

Although successful, the RPA tools did not create a revolution in work, although it allowed a lot of inefficient processes in finance, logistics, order processing, ... to be streamlined and undergo a digital transformation. The RPA market is expected to grow to expected to?$11B by 2027 . Even Microsoft saw the potential, added Power Automate to its arsenal of tools, and is extremely successful with it.

In the meantime, RPA tools evolved, incorporating more artificial intelligence and integrating conversational abilities through another burgeoning technology known as chatbots. These tools seemed promising for customer service applications, but they could only replace the most rudimentary aspects of human customer interactions. While they were adept at retrieving information, they fell short in casual conversations, empathizing with emotions, and comprehending or making jokes. One notable example was the Google Duplex call that aimed to schedule an appointment.

AI's Big Bang in work came on November 30, 2022, when OpenAI released ChatGPT to the world . It was a sort of iPhone moment because the whole world, not just the process engineers in larger companies, suddenly saw the potential of generative AI.

In the meetup "AI & the Future of Work", Elke explained how human resources should deal with the arrival of superpowerful tools that can match humans in knowledge work.

Yes, she introduced us to the Kubler-Ross change cycle (which reminds nerds like me of the Gartner Cycle) with its 5 Stages of Grief. People are angry at first about this new technology, and even want to ban it or at least ignore it. They are also afraid that it can take their jobs. As they realise that this can't be undone, they will look for ways to use this and may even get very excited about it.

Elke illustrated this by the Stevie story, where people felt sorry about the R2D2-like security robot in Washington DC that accidentally landed into a fountain of a shopping center it had to patrol. Meanwhile, delivery robots are being vandalised because people hate them and don't want them to take their job...

Although generative AI tools like ChatGPT or Midjourney look less intrusive, their impact may be many times bigger, particularly on knowledge work.

A few examples:

• copywriters who produce marketing material (often just called copy) use these tools to come up with results faster, with different angles, wordings and styles. There are now hundreds of GPT-based copywriting tools, like copy.ai, Jasper.ai, WordTune, etc. They'll get you over your writer's block by suggestions several ideas, rewrite old content, and generate eye-catching content titles.
• Training organizations can now delegate monotonous tasks such as monitoring student progress, administering interim exams, and recommending assignments to robots. For instance, the remarkable Khan Academy intends to implement this approach to scale its personalized training and tutoring services.
• Developers can use programming copilots like GitHub's Copilot X (or CodeGPT in Visual Studio Code) to speed up their programming, get suggestions, find errors, and write automated tests.
• Every popular tool like Photoshop , Notion , Office , ... gets AI-infused, the incredibly popular Canva tool already had some text-to-image features for some time: just describe what has to be in the image you want, and voila, out comes a suggestion in seconds...

The potential applications of generative AI tools are endless, and we just pointed out some text-to-text or text-to-code tools (there are also hundreds of very popular text-to-image tools, text-to-music, text-to-video, ...). The new GPT-4 can also do image-to-anything generation, which opens up a whole new set of potential applications. Looks like the sky is the limit.

Similar to RPA and early chatbots, generative AI can effortlessly handle mundane, repetitive tasks. However, GPT-based tools possess the capability to extend their functionality significantly beyond these limitations.

In fact, these tools excel in engaging in casual conversations (chitchat), interpreting emotions, sharing jokes, and persuading people that their content is generated by humans. Be cautious of fabricated news articles, images, and videos, as well as highly convincing phishing emails. Did you see the fake photograph of the alleged arrest of Donald Trump ? Fact-checking and critical thinking will emerge as increasingly vital skills in the future.

Loose ends and grey zones        

Unfortunately, our 2-hour meetup could not answer all of the questions from our participants, so there is plenty of content for more meetups. Perhaps we should do a biweekly event, because there are new developments, tools, use cases, insights, reports, etc almost every day. To stay informed, check out the weekly newsletters at Synthedia , AI Weekly or GenerativeAI.net .

The importance of addressing ethical concerns, safeguarding security and privacy, upholding intellectual property and copyright, combatting false information, and understanding the roles of IT and governance when utilizing AI tools in the workplace cannot be overstated. To prevent potential misuse of these tools by employees, which may lead to complications for companies, it is essential to establish and enforce appropriate policies and guidelines.

Do not hesitate to contact us with your questions and remarks,

Patrick Van Renterghem

PS: we already organised a more technical meetup about Generative AI with Sam Hendrickx and Michiel Vandendriessche of the Raccoons Group . You can find the recording of that meetup (in English) on our Youtube channel .