Geertrui Mieke De Ketelaere is the leading lady of the AI world. She calls herself an engineer with a twist. With ‘Wanted: Human-AI Translators’, she has written a bestseller about artificial intelligence (AI), which was promptly translated into French and English, with a German translation due soon. “It’s my mission to show what AI can and can’t do.”
What do you think AI’s biggest achievement is?
Mieke De Ketelaere: “AI can be an extension of the human brain. Nature still has so many secrets. AI can help us unlock these. To give a playful example: show AI a picture of a retina and it will tell you whether it is a man’s or woman’s. Not even the best ophthalmologist in the world can do this. So, AI can remove certain uncertainties in our world.”
What are AI’s biggest challenges today?
“There is no framework. You could compare it with the car in the fifties. Not energy efficient. Safety left much to be desired. And yet, we did not ban the car. We improved the technology and introduced regulations. We worked on the engine, made wearing a seatbelt compulsory and brought in the driving licence.”
“Just like the car, AI is a means to freedom, but there are major issues that need to be addressed. It’s not energy efficient and it excludes people based on gender or ethnicity, because it copies our prejudices.
AI has to be given its place in the new economy, in the discussion about sustainability, taking into account human rights. It has to reflect our hard-won social advances. That is the biggest challenge.”
We have such an inflated view of our own brain that we, superhumans, forget all about the intelligence of nature.
Did you write the book to tackle these challenges?
“The purpose of the book was to find a balance in the debate between the evangelists and critics. AI is a tool that can make our lives easier and remove uncertainties. It’s not a religion. It did not originate in nature, it’s something we have created ourselves. We can control it, but we do have to agree on where we want to take it.”
“Most people don’t lose any sleep over AI. I want to show them that AI is everywhere. It’s a hidden technology embedded in our devices. In the early days of the car, when the first collisions started to happen, people could see with their own eyes that there was a need for regulations. AI is invisible, so with my book, I want to raise people’s awareness of its impact.”
In your book, you make a case for AI translators in companies and organisations who can help us better understand AI. Who are these people?
“Think of them as the general practitioners of AI. Generalists who have an all-round knowledge of the field and refer to specialists where necessary. Unfortunately, we don’t cultivate these kinds of profiles at universities or in companies, where we work in silos. And because of this focus on specialisation, we are not investing enough in bridge builders who connect sectors and fields.”
“Everyone involved in AI has to be willing to translate their field. The technical wizards focus on operation. The businesspeople think about value and associated profit. And, legal deals with liability. They should humbly move beyond their fields and think about the social impact, ethics, and moral framework together. This is where the AI translator comes in.”
You talk about a ‘Hippocratic oath’ for developers, so that, like doctors, they follow a moral code.
“Engineers are, of course, not solely responsible for AI, but they play a vital role. An engineer wants to build, in the same way that a doctor wants to operate. And yet, it’s also his or her duty to think critically. We should start more often from the problem we want to solve.”
“Take Siri. ‘Can you switch on the lights?’ As an engineer, I get a kick out of being able to operate the lights with my voice. Except, it takes energy to do so, because not only does Siri have to ‘listen’ continuously in stand-by mode, it also takes energy to code my voice command into action via a speech algorithm. Sometimes, it is more economical to just leave the light on for a while, or simply turn it on manually. This system does, however, benefit my friend with MS.”
“We have to move away from the ‘toys for boys’ notion and use AI where it makes a difference. For instance, on principle, I don’t use Waze because it fails to capture the complexity of the world. While the system does take into account school hours now, these don’t apply in exam times. But Waze continues to route vehicles to roads with cycling children. Irresponsible.”
The purpose of technology is to give people more freedom.
What drew you to AI?
“My laziness (laughs). I skipped lectures and often copied syllabuses. That took too long, I found, so I built a pneumatic arm that opens the cover and turns pages. It took me longer to build the machine than it would have taken me to do the actual copying (laughs). But it does go to the heart of my belief: the purpose of technology is to give people more freedom. Systems that are capable of independent learning fit in perfectly with this philosophy.”
I read in an interview in Trends that during the coronavirus pandemic you became disappointed in AI
“I became conflicted. Right before the coronavirus pandemic broke out, I had all but finished my book, with only the last chapter left to write. This was supposed to be an account of my visit to the Consumer Electronics Show in Las Vegas, the trade show where everybody who is anybody in the tech world exhibits the latest gadgets. I, myself, also gave a talk there, about decentralised AI. Instead of sending data across the oceans, as Google and Facebook do, we, in Europe, want to activate algorithms locally. That’s more energy efficient, better for privacy, provides faster insights and works in places with limited connectivity.”
“I was super excited about the impact on privacy and energy efficiency, but I only got questions about speed and connectivity. In other words: about how to make even more money by serving more people faster. Stunned, I walked off the stage. As I wandered through the exhibition, I suddenly saw this emphasis on profit, profit, profit everywhere. What about people and planet? When I tried to write my concluding chapter in my hotel room, I had a mental block. I returned home empty-handed.”
We have to move away from the ‘toys for boys’ notion and use AI where it makes a difference.
Then the coronavirus struck, and the world changed.
“I thought to myself: Okay, this is it, this is AI’s opportunity to show the world what it can do. Thanks to AI and the experts, this will be all over in a couple of weeks. Things turned out differently. It was not technology or data that failed us, but people and processes. Experts contradicted each other. If they don’t agree, you can’t train an AI system. There were no agreements on the sharing of data.”
“We went into lockdown. Like so many of us, I sought refuge in nature. I watched starlings swarm graciously in the sky and asked myself why we can’t even prevent two drones from colliding into each other in the lab. I began to immerse myself in the intelligence of nature. Scientists in this sector have to fight hard for their funding, while we, in AI, receive vast amounts of money for brain research. We have such an inflated view of our own brain that we, superhumans, forget all about the intelligence of nature. This connection, about the other forms of intelligence, was missing in my book. My closing chapter developed in nature.”
Remarkably, you also advocate the development of soft skills in education. This in contrast to the constant focus on the STEM debate (science, technology, engineering, mathematics).
“I’ve noticed that communication skills are disappearing. Consulting and listening are considered too weak. It feels safer to get together with like-minded people, but this doesn’t bring about real progress.
I am an engineer with a twist. I love to philosophise and like to think outside my own field. Unfortunately, the act of debating issues and disagreeing on things is not cultivated enough.”
Your book has since been published in three languages. AI tools are widely used in translation work. And yet, you decided not to use it?
“No, I deliberately didn’t want to. One of the translations agencies I had contacted for a test translation, did use AI. I noticed it straightaway. AI doesn’t understand context, which is why it makes mistakes that a human would not. It still feels too artificial. Sure, a song or painting made by AI can be very funky or ingenious. But it lacks human soul. There is something ... creepy about it. It also raises interesting questions: who owns the copyright, is art made by AI even art etc. That fascinates me.”
“I have been asked to curate the Festival of the Future in 2024. During the festival, we will explore experience in the future: how do we view fashion, what – if any – is the role of theatre, and what about our sensory perceptions? This will generate interesting discussions. I want to show what AI can, but also what it can’t do.”
Geertrui Mieke De Ketelaere is director AI at Imec (IDLab), where she helps shape the AI strategy. She was a nominee for the Trends magazine ICT Woman of the Year 2018 award. She is also Adjunct Professor at the Vlerick Business School. Over the last 25 years, she has worked for multinationals such as Microsoft, IBM and SAP in data and analysis.
'Wanted: Human-AI Translators' by Geertrui Mieke De Ketelaere, Pelckmans