ChatGPT is gaining popularity thanks to its extensive training on massive amounts of data, which enables it to provide valuable insights through chat interactions. Its flexibility and ability to generate high accuracies of interaction response make it a versatile tool for various industries. We also see this trend extending to the automotive industry. General Motors, for example, recently announced its plans to incorporate ChatGPT into its vehicles to improve the driving experience, which would allow drivers to communicate with their cars and receive personalized recommendations and assistance.
Fueling Our AI Power
VicOne recognizes the benefits of the generative pre-trained transformer (GPT) family of language models. In fact, we have utilized the strengths of a GPT model to advance our AI capabilities.
Our vast threat database, which has been accumulating threat intelligence for over 30 years and includes over 140 billion blocked threats, has enabled us to train and fine-tune the GPT model with our unique automotive cyberthreat expertise and research insights. Consequently, the model has the ability to learn attack behaviors and detect threats with a high degree of accuracy. It can analyze an attack flow by thinking like an attacker, enabling it to detect even the most sophisticated attacks in the automotive industry.
Furthermore, this GPT model is trained with our unique Automotive Attack Mapping (inspired by MITRE ATT&CK®), which enables it to encompass IT, OT, and automotive-specific tactics, techniques, and procedures (TTPs). Aside from common TTPs, the model also learns 19 additional threat techniques that are specific to connected cars, such as advanced driver-assistance system (ADAS) sensor attack, exploit via Unified Diagnostic Services (UDS), and exploit electronic control unit (ECU) for lateral movement. This exceptional knowledge is especially valuable in the automotive industry, where cyberthreats are becoming increasingly prevalent.
By integrating a GPT model that is focused on the automotive industry and trained with data related to automotive cybersecurity, we can reduce our machine learning operations (MLOps) pipeline’s data analysis process time by 60%. This GPT model helps facilitate certain tasks, such as data cleaning, reducing training dataset sizes, and correlating anomaly events using GPT’s natural language processing (NLP) embedding model. It also greatly improves the average false positive rate before automotive cybersecurity threat expert fine-tuning.
With a customized GPT model for automotive cybersecurity, combined with the analytics engine behind our cloud-based XDR platform, xNexus, we can help security analysts in vehicle security operations centers (VSOCs) determine the root causes of issues in the connected car ecosystem faster, better understand the attack context of different ECUs, and even detect potential threats before the attack chain is fully executed. This can expedite investigations and enable analysts to respond to threats preemptively.
Staying Ahead of Threats
As cars become more advanced with AI and their surrounding ecosystem becomes increasingly complex, the risk of sophisticated and innovative attack behaviors also rises. To address these challenges, VicOne leverages its existing strengths, including the 102-Day Early Protection. Combined with the power of our automotive cybersecurity-knowledgeable GPT model, we can analyze abnormal behaviors more accurately and we can detect harmful behaviors more quickly. With our reliable and efficient solutions, we are dedicated to addressing cyberthreats in the automotive industry, safeguarding the safety and security of connected vehicles and critical systems.
Read our other blog entry to learn how AI models like ChatGPT can influence automotive cybersecurity.
