Symbolic AI and the Future of Autonomy

Redefining Intelligence with Understanding

At the core of ototo’s solution for autonomous driving lies an unusual kind of artificial intelligence - one not built solely on data, but on deep understanding. Symbolic AI, an established, if lesser-known, branch of artificial intelligence, is now proving to be the missing piece in the puzzle of full vehicle autonomy.

Symbolic AI mirrors the way humans reason. Instead of just detecting patterns in massive data sets, it uses logic, structure, and causality - just like we do. With its combination of the power of AI and the beauty of formally modeling human domain expertise.

Why Symbolic AI Captivated Me from the Start

My fascination with Symbolic AI began decades ago, before it was even commonly called by that name. As a mathematician working on algorithmic problems, I have always looked for solutions that didn’t just work, but that made sense to us humans. Symbolic AI has helped me build high-performance algorithms in different fields - including the world’s first in-camera red-eye correction algorithm which was deployed in over 120,000,000 digital cameras. With Symbolic AI and over 14 patents in the field, my career has been defined by solving problems others deemed too complex, from fraud detection to healthcare, identifying everything from suspicious banking sessions to embryo genetics.

Ron Maor, co-founder of ototo, mathematician and serial entrepreneur.

The Real-World Driving Challenge

This ability to abstract human-like reasoning into an algorithm is exactly what autonomous vehicles have been missing. While classic AI approaches have made incredible progress in detecting objects, they struggle with the crucial next step, to make sense of the road. Especially in edge, complex, or dynamic situations, advances in AI have not translated into driving algorithms that can face the real world. These “long-tail” road scenarios, are where today’s black-box systems fail. And these can arise without notice, in urban driving and even on the highway.

This ability to abstract human-like reasoning into an algorithm is exactly what autonomous vehicles have been missing.

To quote a keynote speaker from the automotive field, “Prediction is the hardest part of autonomous driving”. Prediction – and the path planning that depends on it are about more than seeing the world. They are about understanding intents, characteristics, interactions, dependencies and anomalies. There is no way to learn all that from data – in mathematical terms, the solution space is just too big.

One End-to-End SAI Foundation Model- Multiple Capabilities

Enter ototo. Our company was founded around the premise that Symbolic AI is uniquely suitable to this hitherto-intractable problem. Our Symbolic AI Foundation Model is rewriting the rules of autonomy. Our model doesn’t just anticipate road users’ behavior, it explains its reasoning step-by-step as part of the system outputs. It adapts in real-time, handles edge cases gracefully, and requires a fraction of the computational load that traditional systems demand.

One Model- Multiple Capabilities

And crucially, it does this in long-tail cases as reliably as at all other times, and it does it fully transparently. Full coverage, and user trust, are major issues in autonomy. People need to see the system handling all situations, and to feel that “it knows what it’s doing” rather than that “it got lucky this time”.

Symbolic AI is the key to all this.

People need to see the system handling all situations, and to feel that “it knows what it’s doing” rather than that “it got lucky this time”. Symbolic AI is the key to all this.

From Early Experiments to a Mission-Driven Company

From 2021, I started playing around with the algorithm, refining it and exploring different applications. After landing on the right market fit, we founded ototo in 2023 with a clear mission: to take a different approach to tackle the core bottleneck in autonomous driving, long-tail road scenarios, by building an end-to-end solution with Symbolic AI at its core. With operations now spanning Germany, UK, and Japan, and growing traction with OEMs and Tier 1s, ototo is leading a quiet revolution in the mobility space.

Looking Ahead

I am excited about what lies ahead, not just for ototo, but for AI as a discipline. There is a shift happening. People are realizing that more data isn’t always the answer. Sometimes, what’s needed isn’t just more data- but a fundamentally different kind of model. One that understands context, adapts to uncertainty, and explains its decisions with clarity. Nowhere is that more critical than in autonomous driving.