blogs
Separating fact and fiction in AI
Since the launch of ChatGPT just over two years ago, artificial intelligence (AI) has been thrust into public consciousness. It has rapidly transformed industries and redefined workflows, yet the line between fact and fiction in AI remains blurred. This is because most of the focus has been on 'Large Language Models', which have been trained on the vast amount of data available 'for free' on the internet. As such, these models are restricted to language, text, and images, and their output can often be incredibly subjective.
To date, we have only scratched the surface of AI’s capabilities. It is the most powerful tool at our disposal since the creation of the Internet and is already being used in some world-changing fields, such as drug discovery, with companies such as Isomorphic Labs leading the way. But what if we could use AI to benefit society even further, such as in engineering? The real value of AI lies in addressing the complex problems humans struggle with.
To do this, we have to take a more focused approach to AI development. If we save all the data from millions of AI-driven simulations within physics over the next few years, we can create more efficient and precise models based on proven scientific facts rather than opinions. These models will enhance human abilities rather than mimic them, potentially unlocking new avenues of innovation. For instance, in engineering, we can create more efficient designs, bring down costs and change how we invent.
The automotive industry is an excellent use case for this more focused AI approach. The West is currently running hard to try and catch up with China regarding electric vehicle (EV) development and adoption. In 2024, China’s largest car manufacturer, Build Your Dreams (BYD), outsold and outproduced the incumbent, Tesla. Aside from game-changing engineering capabilities, China also dominates supply chains, holding end-to-end control over the means of production. In particular, its access to raw materials offers the nation a significant advantage. If other automotive companies hope to compete, the cost of producing EVs must be brought down dramatically. This is where AI can play a key role.
We’re at a point where human engineers cannot produce the same number of designs as AI. AI can already explore hundreds of thousands of prototype designs in just a few days and generate design options for manufacturers that can reduce the cost of key components.
Taking an EV’s motor system as an example, with just a 10% cost saving, the average real-world saving for Original Equipment Manufacturers (OEMs) would be in the region of £1bn for large OEMs over the production run of a single motor system. For average-sized OEMs, the savings are around £100m. On top of this, the increased speed of design saves time and reduces the associated costs. Combined with the performance improvements, the cost savings across the board could be transformative for OEMs.
We are at a critical inflection point in the AI revolution. The true value of the technology lies in its application to the objective rather than the subjective. Although Generative AI has developed at a pace and impressively mimics human capabilities, applications in the sciences will see humanity reach new heights and transform the way we invent.
