“We take the robot out of the human.” That is the promise of Robotic Process Automation (RPA) consultants and developers. Now that technology giants such as Microsoft and Meta are rapidly making substantial investments in making their products and services smarter, this is anything but an empty promise.
The era in which artificial intelligence is part of every job is now over. The IT manager would do well to take the enormous adoption rate of writing robot ChatGPT and other groundbreaking tools seriously, with all the pitfalls that this entails. It is important to proactively guide the organization in the process in which AI will soon end up in every workplace.
If Robotic Process Automation manages to keep its promise, the question quickly arises: what exactly will remain when the robot has been removed from the human? This question is moving up the agenda of IT managers. Just about every knowledge worker has or will have an AI buddy. The arrival of ChatGPT is a tipping point for artificial intelligence.
Two months after the launch of ChatGPT, more than 100 million people were already using the smart chatbot. By comparison, TikTok took nine months to reach the same numbers. AI is increasingly moving from doing work in the background to the foreground. Anyone who now unconsciously uses AI behind the scenes will soon be consciously working with artificial intelligence on a daily basis.
Smart, smarter, smartest
In collaboration with OpenAI, Microsoft is fully committed to slimming down. ChatGPT writes LinkedIn messages and Microsoft Valves makes custom sales proposals. Soon Microsoft Teams will automatically generate agenda proposals and minutes of meeting sessions. And GitHub’s Co-Pilot helps with programming. AI applications such as ChatGPT, Dall-E, Midjourney, Stable Diffusion and RunawayML accelerate creative processes. Anyone can produce a poem, painting, film, marketing campaign or program code, as long as you enter the correct words in the text bar. Creativity is democratizing and it also shows how big the impact of the new generation of AI applications is.
Andrej Karpote, inventor of the Tesla autopilot, says he already has 80 percent of his code written by the intelligent GitHub Copilot. Standard programming is making way for so-called ‘prompting and editing’; fine-tuning the codes delivered by GitHub Copilot. This is a good example of how smart knowledge workers are quickly but critically embracing artificial intelligence.
Getting people and artificial intelligence to work well together is a tough task. This requires training, practice and sometimes even retraining of employees in the company. The emergence of generative AI also certainly entails risks. After all, cyber criminals also get their hands on these tools and competitors are also doing their best to strengthen their position with AI tools. However, we must not forget that many tools are still far from functioning perfectly. Certainly the reproduction of prejudices is still a hot topic.
Ask an artificial image maker like Dall-E 2 for a lawyer and the app only returns white men. Everything that is produced has its origins in human-written texts, works of art and product designs. So in a sense, AI is holding up a rather confronting mirror to us. In addition, there are plenty of discussions about issues such as copyright and IP. And although texts often come across as convincing, a writing robot like ChatGPT can also spout sheer nonsense. In short, the role of humans is certainly not over yet.
In this way, artificial intelligence is mainly becoming a digital assistant, co-pilot or sparring partner. In the ideal situation, we can confidently leave routine tasks to AI. What remains is a person who can flourish in creative and problem-solving issues and who can excel in fine-tuning. But the so-called ‘happy flow’ is not a given. A good interaction between humans and artificial intelligence is becoming just as important and complex as cooperation between people. It is indeed yet another extra job on the plate of the IT manager, but there is a lot to be gained by proactively learning to deal with the smartening of the acceleration in digitization.
And therefore also Robotic Process Automation. Do not immediately make the experiments too large, but keep it small. And keep reflecting on the interaction between humans and AI. In this way, the technology officer is getting a better answer to the question of what remains for humans in interaction with the robot.
Thijs Pepping is a trend analyst at the Research Institute for New Technology (VINT) of IT service provider Sogeti