Who is Yoshua Bengio – And Why He’s Scared of What He Created
There are several key figures behind AI, and Yoshua Bengio is definitely one to be named. This Canadian computer scientist is one of the pioneers of deep learning. But who is he on a deeper level, what are the things he’s done, and how does he currently feel about AI?

Dianne
Updated January 29, 2024

Yoshua Bengio smiling in a grey room, made with Midjourney
Reading Time: 11 minutes
Yoshua Bengio is an influential figure in the AI industry who set out the collective efforts in furthering AI. His sentiments regarding the use of AI are widely significant in his field. He steered the direction and progress of fast-paced research on intelligent machines.
Needless to say, although born in the 1960s, his contributions are as valuable as those of the giants who spearheaded the artificial neural network back in the 1950s to benefit humankind. Their work powered today’s deep machine learning so sophisticated that there’s now a possibility that AI can someday top human intelligence.
So who is he? What has he done? And what is he currently up to? Here’s everything to know about Yoshua Bengio.
Yoshua in his Earlier Years
Yoshua was born to a Jewish family on March 5, 1964, in France, where he spent his childhood. Unlike other children who took computers for granted, Yoshua developed a keen interest in computing for a much deeper reason and started programming at age 11. For him, it wasn’t just a hobby, but a starting point for what he needed to do to achieve the ideal world he had imagined.
As a child, he envisioned and drew his inspiration for a high-tech future from the worlds as portrayed by science fiction authors such as Arthur Clarke, Isaac Asimov, and Ray Bradbury.
Using Apple II and Atari 800, Yoshua and his brother Samy Bengio went on experimenting with machine language until they reached their teenage years and moved to the city of Montreal.
His Educational Journey
Yoshua pursued his scientific aspirations and received his Bachelor of Science in Computer Engineering from McGill University in 1986. It was only during that time that he learned the concept of neural networks, which were invented to mimic the human brain through computational systems.
Given that he’s been fascinated and enthusiastic about sci-fi ever since he was young, Yoshua Bengio took a Ph.D. in computer science to further explore the powers of technology, which he completed in 1991 at the same university.
After his graduation, with all the knowledge he had, he became a postdoctoral fellow at Massachusetts Institute of Technology (MIT) in sequential data and AT&T Bell Laboratories in vision algorithms. He also became a faculty member at the University of Montreal in 1993.
Founding Mila
Within the same year Yoshua became a professor, an AI research institute at the heart of Quebec was born.
Yoshua Bengio founded the Montreal Institute for Learning Algorithms (Mila) in 1993, which is currently the largest academic research center for deep learning. This giant research lab was established with a mission to become the global hub of scientific breakthroughs and foster innovations in technology with a special focus on AI.

In 2017, this Quebec artificial intelligence institute expanded its operation. Now, it brings together a community of researchers, scientists, professors, students, and tech enthusiasts in this ever-evolving field for the benefit of all, though that’s not quite where it seems to be going at the moment given all the wild advancements in AI.
Today, the Mila community is contributing to various areas of AI research, not only to advance this technology but also to ensure that it is good for society. The efforts aim to turn Yoshua’s ideal visions for a safe, high-tech world into a future reality.
Not only is Yoshua a director of the research institute that he founded, but he is also the scientific director of IVADO and a co-founder of Element AI in 2016 (now acquired by ServiceNow). He also co-headed the Learning in Machines & Brains program at the Canadian Institute for Advanced Research (CIFAR) with Yann LeCun, a program director until March 2022.
He also became a fellow of the Royal Society of both Canada and London, the Canada Research Chair on Statistical Learning Algorithms, and many more. He is an active contributor to his field.
Neural Networks and Deep Learning
In 1998, the idea of document recognition was introduced—thanks to the groundbreaking paper titled Gradient-Based Learning Applied to Document Recognition by Yoshua Bengio and Yann Lecun, and two other researchers, Leon Bottou and Patrick Haffner.
Their paper proposes a new learning paradigm called graph transformer networks (GTN), which analyzes a piece of document as a whole by treating the text characters, images, overall layout, and other elements as nodes, and their relationships as edges in a graph to understand what’s in the document, what they are for, and what needs to be done.
Because of that, we have text and image scanners today, but that’s not the only application of this neural network architecture; it’s also used in fraud detection, social network analysis, summarization, and even question answering (take MathGPT, for example).
Another famous paper Yoshua published in 2000 further enhanced the computer’s level of understanding of the human language. This paper, A Neural Probabilistic Language Model, led to the development of sophisticated AI technologies that we widely use today. Some of them are the predictive text on our phones, autocomplete suggestions, autocorrection (though sometimes, we hate it), and language translation.
Yoshua Bengio is a pioneer in the realm of artificial intelligence for a big reason. His deep learning and neural network discoveries have brought us to where we are today. Although he wasn’t the only catalyst for these technological advances, he’s certainly among the few most influential figures in machine learning.
Let’s move forward to how he became one of the key movers in turning a once science fiction into a modern-day reality.
Becoming One of the “Three Musketeers” of Deep Learning
It all started when Yoshua Bengio met Yann LeCun for the first time after his graduation. Together, they started their journey in computing through a simple collaboration on a project based on Yoshua’s Ph.D. thesis, which revolved around a system for handwriting analysis. AT&T then used the system to automate paper check processing, transforming the banking industry in the process.
Fast forwarding to 2015, Yoshua and Yann published their research on deep learning with Geoffrey Hinton, whose works focused on the nature of human intelligence greatly inspired Yoshua Bengio to unravel the possibilities and extent of intelligence with respect to “lifeless” machines.
In 2018, the three shared the victory in the Association for Computing Machinery (ACM) as they received the Nobel Prize of the Turing Award. Yoshua was recognized for being one of the first to combine neural networks with probabilistic models in natural language processing (NLP), leading to the emergence of speech recognition systems.
With Yann LeCun and Geoffrey Hinton, Yoshua holds the position of being one of the three prominent figures in their field, earning them the connotation of “three musketeers” in the world of science and technology, not in a war, although they’re already kind of sensing a “battle” ahead.
His Work on Generative Adversarial Networks
Going back to 2014, Bengio, with his Ph.D. student, Ian Goodfellow, had another breakthrough when they invented the concept of generative adversarial networks that use unsupervised learning. So how does it work?
To simplify, there are two competing networks, in which one acts as a generator while the other as a discriminator. The generator AI model creates outputs based on input or prompts, and the discriminator judges the quality of the output generated by the other network based on how real data—it is pre-trained on—looks like, and then the generator applies those feedback to improve its output and the process repeats until the discriminator network can no longer distinguish the work of AI from real human output.
So basically, it’s an AI versus AI duel in a way that one aims to create as realistic output as possible and trick the other into thinking that it is human-generated. This applies now to famous AI art and image generators like Dall-E and Midjourney. So if you’re fond of and a heavy user of AI image generators, you know who you should be thanking now.
Awards, Distinctions, and Publications
Along with Geoffrey Hinton and Yann Lecun, who’ve made massive contributions to the field of AI, Yoshua Bengio has become one of the thought leaders in artificial intelligence who received the “Nobel Prize of Computing”, which is known as the ACM A.M Turing Award in 2018.
Apart from the prestigious A.M. Turing Award in 2018, do you know that Yoshua Bengio was the most cited and third most influential computer scientist in the world in 2022? Well, these are not the only things he’s known for.
Here are some of his other achievements:
- Princess of Asturias Award, 2022
- Killam Prize, 2019
- IEEE CIS Neural Networks Pioneer Award, 2019
- Lifetime Achievement Award, 2018
- Officer of the Order of Canada, 2017
- Marie-Victorin Quebec Prize, 2017
His select published research includes just some of:
- Deep Learning (Adaptive Computation and Machine Learning) (2016)
- Neural Machine Translation by Jointly Learning to Align and Translate (2015)
- Deep Learning (2015)
- Generative Adversarial Networks (2014)
- Advances in Neural Information Processing Systems (2009)
- Greedy Layer-Wise Training of Deep Networks (2007)
- A Neural Probabilistic Language Model (2003)
- High Quality Document Image Compression with DjVu (1998)
- Gradient-Based Learning Applied to Document Recognition (1998)
Contemplating the Aftermaths of AI in the Wrong Hands
Despite all the wins and being in the limelight, there’s something dark about AI that Yoshua dreads. Well, it’s not really the AI, but the “bad actors” who would use AI to carry out disastrous plans.
In an interview with BBC in May of 2023, Yoshua opened up that he felt “lost” over his life’s work. Now that AI technologies are becoming much more sophisticated and more powerful, he’s growing more anxious about the potential dangers they could bring about to humanity when they fall into the wrong hands.
“It might be military, it might be terrorists, it might be somebody very angry, psychotic. And so if it's easy to program these AI systems to ask them to do something very bad, this could be very dangerous.” Yoshua mentioned.
Particularly, he worries about China’s use of AI, which he solemnly expressed during his interview with Bloomberg. With China’s global surveillance integrated with facial recognition that can be used for manipulation and its latest burning issue regarding its military use of the Baidu chatbot, the AI sector was shaken to its core by the possibility of China’s techno-authoritarianism. If there’s one thing for sure: no one wants to live in a sci-fi dystopia.
Yoshua well understands how fast this cutting-edge technology he helped propel moves, so like Geoffrey who believes AI could become smarter than us and is now filled with regrets, the foreboding phenomena in AI took a huge toll on him as his realization that the double-edged nature of AI, power and vulnerabilities, is taken advantage of and abused dawned on him.
But will he be able to stop the far-reaching impact of the AI threat before it spreads like wildfire, and potentially leads to human extinction? Most likely not. The beast has been unleashed, and it’s only a matter of time until we see its many positive and negative impacts.
Advocating for Responsible Development and Safe Use of AI
Foreseeing the probability of autonomous weapons, widespread misinformation, especially in the coming US election, and all the other existential risks associated with the misuse of AI, Yoshua Bengio decided to take some actions to revert the imaginable catastrophe of what he’s created. Fortunately, he’s not alone in this.
With Geoffrey Hinton, Sam Altman, and many other AI scientists and notable figures like Bill Gates, Yoshua signed the Statement on AI Risk and, recently, an open letter that aims to slow down the development of giant AI systems that pass the Turing test through a six-month break.

He also actively contributes to the Montreal Declaration for the Responsible Development of Artificial Intelligence. This framework for AI promotes ethical deployment, better regulation, more hands-on involvement of the government in AI product registration, auditing, and tracking, more socially responsible development of AI solutions, and stronger adherence of AI systems to human moral code of conduct.
These are huge works, indeed. But despite that “anything that can go wrong will go wrong,” as stated in Murphy’s Law, Yoshua Bengio is determined to reverse the threats of AI and save not only the future of his works but also the future generations of humanity.
He’s an incredible figure who has made an impact on the next few decades for sure. While hesitant about the future of what he’s created, at this point, we can only hope for the best and wish things unfold well.
Want to Learn Even More?
If you enjoyed this article, subscribe to our free newsletter where we share tips & tricks on how to use tech & AI to grow and optimize your business, career, and life.