AI and Trust

Trusting AI is a broad question, spanning ethics, safety, accuracy and more.

Today at the AI Brainport Summit, Djoni de Vos noted that for many clients, accuracy is still the first concern. Anything less than 99% won’t do. May I add, especially in fields like engineering or design, where answers must be precise.

But what happens when disruption comes from large language models, systems that are not deterministic by nature?

In creative or exploratory tasks, that flexibility is a strength. Yet in specialized fields like CAD, architecture, or simulation, it seems that companies like Dassault Systems are betting on domain-specific AI, shape-based or physics-aware models, to maintain their edge.

So, as general-purpose LLMs evolve but may reach a plateau in training data, while specialized models deepen, will the future belong to domain-specific systems or better LLMs?

(On the pictures, fun demo with AI-powered robots… which had to be escorted out at the end of their conversation.)

#AI #AITrust #FutureofWork