Our paper on "Certifiable AI in Aviation", which was largely developed at the Institute for AI Safety and Security, received the "Best of Session" award at the Digital Avionics Systems Conference (DASC) in San Diego! In this article, we show how important artificial intelligence (AI) is becoming in technology and what safety issues arise from this, especially in aviation, where particularly strict certification requirements apply. We have therefore presented an open-source implementation of SafetyNets, which uses special sparsely populated look-up tables to correct errors in neural networks. By combining these tables with neural networks, we create a certifiable system that offers 100% reliability and fast error correction in real time. A big thank you to our dedicated team for this outstanding achievement - together we are making the future of aviation safer!
