Wat mengs du? – AI makes choices for us: what do we still want to decide for ourselves?
Switch to French for original article
Listen to this article
Booking a restaurant, preparing a legal file, helping the children with their homework: in the future, AI agents could do all of this for us from a single screen. Serge Linckels, director of the Digital Learning Hub, sees this as an opportunity to make knowledge more widely accessible – provided we retain control over our choices, as he explains in his opinion piece.
This free access is available to you at no cost. If you'd like to read our editorial team's articles and support quality journalism, subscribe now.
Recently, I wanted to book a table online at a vegetarian restaurant in Luxembourg City that's open on Monday evenings. Like you, I browsed websites, scrolled through the list of pages suggested by Google, explored social media and laughed at the cat videos that popped up, before asking myself what I'd actually set out to find in the first place. I often find myself getting lost in the fantastical clouds of the web and social media, losing track of time and forgetting the original purpose of my search. It's a real challenge to resist the entertaining and frivolous content that grabs my attention.
Here, I'd like to share my vision of how the world of screens is set to change with advances in AI. In the future, your AI assistant – such as Mistral's Vibe or OpenAI's ChatGPT – will perform much more sophisticated tasks. It will have got to know you through your interactions and will liaise with AI agents from restaurant booking services or home delivery companies, for example, rather than consulting websites or relying solely on its model's knowledge.
More specifically, your AI assistant will make a booking directly based on your tastes and preferences, send a confirmation message, add the event to your calendar and send the restaurant's details to your car's sat-nav system. The AI agent will carry out an entire task on your behalf and then return to you to ask you to make a choice or confirm. In other words, the AI assistant switches to execution mode: this is the principle of agent-based AI.
"Entrusting tasks [to AI agents] does not mean relinquishing our choices to them. It is up to us to decide what we wish to delegate and what we want to keep in our own hands."
Serge Linckels
This new world will consist of a network of AI agents and will offer enormous benefits to users. Gone are the days of browsing the web and getting lost whilst searching for information. You'll access this network of agents via a single screen: your favourite AI assistant. You'll continue to use your local apps, but will go through the assistant to search for information or carry out complex tasks. This screen will be simple: it will essentially consist of a text field where you'll state your intention. We'll acquire new skills: the ability to express ourselves via a single screen and to develop a critical mindset towards AI.
Whilst the example of booking a restaurant is admittedly limited, here are some more promising areas. A network of medical AI agents would possess an incredible wealth of knowledge about diseases and their treatments to advise your GP. A collective of legal AI agents would have access to the full body of laws and case law, enabling you to defend your interests in court on a level playing field and at an affordable cost. Your children would be supported in their studies by a team of AI agents who would apply the teaching methods best suited to their needs.
This approach will increase our dependence on AI, but could also democratise access to knowledge. Surrounded by a team of AI agents at your service, you will discover unexpected possibilities and new perspectives, rather than being reduced to a profile. The diversity of the agents and their mutual oversight could foster greater objectivity and more transparent treatment.
About the author
-
Serge Linckels is a computer engineer and holds a PhD in Internet Technologies and Systems from the Hasso Plattner Institute, where his research focused on artificial intelligence. He taught for over twenty years at the Guillaume Kroll High School and was an associate lecturer at the University of Luxembourg. The author or co-author of some thirty scientific publications and three books, he gives lectures on AI and its impact on learning and society.
Serge Linckels heads the Digital Learning Hub, a public adult education institute in Luxembourg. The institute offers a range of high-quality modular training courses that enable adults to develop their digital skills throughout their lives. Mr Linckels also runs 42 Luxembourg, a free programming school based on project-based learning, as well as three academies dedicated to artificial intelligence, cybersecurity and quantum computing. Through Elements of AI Luxembourg, the Digital Learning Hub offers a free introduction to artificial intelligence open to everyone.
However, some AI agents might promote a product or an ideology. Their interactions must make it possible to highlight differences of opinion and identify instances of manipulation, or even to exclude agents from the trusted network. An ethical framework must enforce a requirement for transparency, and mediation mechanisms will enable conflicts between agents to be resolved.
A single screen must not become a single source. Users must be able to see the sources consulted, compare different viewpoints and access alternatives to those selected by their assistant. This transparency should also enable them to identify the AI agents involved and detect malicious components, so that they can choose which AI agents they trust. The aim will be to provide users with greater protection against inappropriate or manipulative content, particularly in the case of children and teenagers.
I have outlined my vision of a new world where AI agents could free us from many tedious tasks and broaden our access to knowledge. But entrusting them with tasks does not mean relinquishing our choices to them. It is up to us to decide what we wish to delegate and what we want to keep in our own hands. Perhaps in this way we will build a better and more peaceful world, one that will make us better human beings.
Wat mengs du? (What do you think?)
-
Once a month, we give a platform to a voice – someone who is an expert in a particular field, whether through their studies, their profession or their personal experience: experts on everyday life, on an illness, on a specific life situation – or simply someone with a clear opinion.
Do you have something to say? Then send us your idea for an opinion piece to journal@journal.lu.