How will AI change software quality in cars?

At the 28th Automotive Electronics Congress in Ludwigsburg, Zohar Fox, co-founder and CEO of Aurora Labs, spoke about the role of artificial intelligence in software development. In his keynote speech ‘How AI will be your trusted advisor for SW reliability and quality’, he showed how AI can improve the reliability and quality of software in the automotive industry.

What are the current challenges in automotive software?

The complexity of today’s automotive software is enormous. An example Fox used in his presentation was the ‘spaghetti picture’ – a metaphor for the many interconnected problems that can arise during development. Fox explained, “Every time we change something, there’s the butterfly effect. Everything shifts and changes, and the integrators especially feel this very strongly.” These problems arise because the modules, which come from different suppliers, can cause unexpected problems when integrated into the overall system.

How can AI improve the reliability and quality of software?

Fox explained how AI can be used to overcome this complexity. One of the most important tasks of AI is to recognize patterns in the vast amounts of data generated during software development. AI can analyze these patterns and make predictions about potential problems before they happen. “With AI we can confirm the unknown,” says Fox. An example of the power of this technology is its ability to recognize and correct errors in real time, significantly improving software quality.

What role does binary analysis play?

A central point in Fox’s presentation was the importance of binary analysis. He explained that Aurora Labs has started using natural language processing (NLP) not only for text, but also for binaries. “At the heart of it all is the binary, the images or executables that actually run on the hardware,” he explained. Analyzing these binaries with AI can provide deep insights into the software that were not possible before. This enables more accurate error detection and correction.

How is AI used in practice?

Fox presented how AURORA LABS has integrated AI into existing systems to increase efficiency. She showed how AI is used to optimize code to save memory and computing power. “In this pull request, we were able to free up about 3.5% of memory by removing excess code,” Fox explains. This type of optimization can be performed daily and leads to a significant improvement in software performance.

What does the future of automotive software look like?

Finally, Fox provided a look at the future, in which AI will play an even more central role. She spoke about working with leading companies such as Infineon to develop AI-based solutions that can be used across the industry. “What we’ve created is a Retrieval Augmented Generation (RAG) system that bridges the gap between AI and binary file analysis,” she proudly explained.