The future of AI – blowin’ in the wind

With the number of articles, events and even TikTok movies about Artificial Intelligence, you might wonder if there is any aspect of AI left to discuss. Well, think again. Attending the Miles Ahead Session on the Future of AI on April 19th has definitely gave us new insights and food for thought. No real surprise if you know the line-up consisted of our very own Maarten Mortier and Dr. Marie Charles, aka the ‘Joan of Arc of 21st Century Medicine’. 

The first Miles Ahead Session took place on Miles Ahead’s home turf, the Sociëteit De Verloren Kost on Ghent’s Coupure canal and drew over 40 tech entrepreneurs and investors, eager to discuss deep tech while enjoying cocktails and finger-licking snacks. Not only did the presentations whet their appetite for food and drinks, but also for investing in AI.

Maarten Mortier confronted the participants with the dilemmas of AI, one of the key ones being: should I invest already in AI? Or should I wait for the technology to mature further? Another fundamental question about AI: what is it really? Do we all mean the same thing when we use the term? Is Deep Learning the same as AI? Or is Deep Learning merely a toolset for AI practitioners? 

Concept Validation

(I never promised you a) rose garden

Just like every new technology, AI is moving along the usual adoption curve, moving from the ‘Peak of Inflated Expectations’ and the ‘Trough of Disillusionment’ to a ‘Plateau of Productivity’. Inflated expectations were certainly there, but now it has fortunately been accepted that AI cannot do everything yet and cannot solve all problems in the world.

However, AI has progressed significantly in recent decades, especially since the volume of available data to learn from increased exponentially, and as new models emerged. One of the most hyped models is the ‘Transformer Model’, a neural network that learns context and meaning by tracking relationships in sequential data. Also called ‘Foundation Models’, these models can apply learnings from one domain (for instance NLP – Natural Language Processing) to another area of research (for instance image recognition) and achieve better results. Transformers are a model that require a lot time and effort, making this technology expensive and reserved for Big Tech only. As Maarten Mortier pointed out, AI research is no longer the domain of academics, as it was at the turn of the century. This drastically disrupts the role of open-source academic work in AI. 

Schools out

Great advancements have been made in AI, not least by the development of new learning models such as supervised learning, unsupervised learning and self-supervised learning. In self-supervised learning, AI questions its own results and by trying to fool itself, becomes more robust and accurate. In Maarten Mortier’s view, self-supervised learning is the future of AI. He also expects a lot from neurosymbolics, which is akin to how a child learns through built-up fundamentals like logic, axioms and knowledge trees. 

The eve of destruction?

Although an enthusiast of AI, Maarten Mortier is not blind to the dangers of the technology. “Humanity is not very good at controlling these types of risk,” he said when discussing Artificial General Intelligence (AGI) — the (as yet) hypothetical ability of an intelligent agent to understand or learn any intellectual task that a human being can. The danger of AGI is that it may merely seem close and, during that period, many people will treat its (non-)existence as a religious or political idea. This will obfuscate further development. And while AGI may put jobs at risk, Artificial Super Intelligence (ASI) may put humanity at risk. 

The best is yet to come

Dr. Marie Charles is a firm believer that technology in general and artificial intelligence in particular can have an impact in healthcare and life science and turn the world into a better place to live. After an extensive international career founding, managing and growing multinational companies and organizations, this Belgian top medical scientist returned home to study cause and effect of micronutrients on our health. She is convinced that using AI for this research will lead to transformational change in the way we look at food and will give rise to a completely new industry.   Her reason for coming back to Belgium for this research? Simple: the best talent on AI is to be found in Belgium. Who are we to disagree?

After this session, it was obvious to all attendees that the ‘trough of disillusionment’ is well behind us when it comes to AI, and expectations are now realistic rather than inflated. The time has come to start capitalizing on all the research that has been done and founders with a purpose can now turn ideas into ventures. With a little help from your friends at Miles Ahead of course. 

The next Miles Ahead Session (May 24th) will focus on InBound marketing and features presentations by Ewout Meyns (Co-founder/CEO of PieSync) and Rory Hope (Head of Content SEO at Hubspot). 

For more information and registration, click here

Other Stories

We're hiring!

We are always looking for the following profiles for our Studio and our first ventures.

Current Opportunities