Amazon Web Services (AWS) for Automotive recently announced how they see generative AI affecting the auto industry—creating software-defined vehicles and control systems, improving vehicle hardware performance, enabling OEMs to create simulations to test driving scenarios, etc.
Similarly, back in April AWS announced Amazon Bedrock and multiple generative AI services and capabilities, indicating their eagerness to move industry forward with these tools while democratizing access for builders.
It's exciting stuff—multifaceted and evolving at a rapid clip. Here we chat with Wendy Bauer, general manager of AWS Automotive & Manufacturing, to dive a little deeper and learn about their new initiative focused on this tool. Take a look...
Smart Industry: What is the AWS Generative AI Innovation Center? What is the rationale for its launch?
Wendy: The AWS Generative AI Innovation Center connects customers with AWS ML science and strategy experts to identify and implement valuable and innovative generative-AI solutions. Our enterprise customers have shown interest in experimenting with generative AI and have turned to AWS for help. The Innovation Center is one part of a broad set of investments to help our customers succeed with generative AI.
The AWS Generative AI Innovation Center is meant to help all of our customers across industries, including auto and manufacturing, select the right use cases to experiment with, improve accuracy in foundation and large-language models, and strategize how to fine-tune and customize these models for use cases.
Smart Industry: How is generative AI changing automaking?
Wendy: Generative AI can uplevel automaking in a multitude of ways, from engineering and performance optimization to in-vehicle experiences and customer service. For example, we envision generative AI optimizing the design of mechanical parts to reduce drag in vehicle designs. When it comes to in-vehicle experiences, it will allow the creation of new concepts such as personal assistants. Generative AI can also be used for synthetic data generation to test applications, especially for data not often included in testing datasets, such as defects or edge cases.
We see generative AI changing automaking as a whole by customers leveraging the tool to help create new material, chip and part designs. This will optimize manufacturing processes and drive down costs.
Another great example is how generative AI can enable faster and higher-quality vehicle software by helping to analyze existing code, identifying problematic code with higher accuracy, and providing intelligent suggestions on how to remediate it. Because potential vulnerabilities can be found and corrected earlier in the application lifecycle, this can lower the cost, time and risk of development for vehicle software.
True adoption is nascent at this point, as with most industries, though we’re already seeing some adoption and experimentation among auto companies, mostly in the customer-service space. They’re working on using generative AI to deliver better customer service by providing quick responses to the most common customer questions.
Smart Industry: What hurdles are there to overcome in this vertical?
Wendy: Cloud will play a critical role in the automotive industry’s ability to successfully adopt generative AI, particularly because generative AI relies on vast amounts of data that encompasses various aspects such as training, model-development, scalability and flexibility. Curating datasets, detailed specifications, and design information can be a big task for automotive companies to manage. This is where AWS plays a crucial role.
Generative AI requires a large amount of computational resources and data, which can be costly and time-consuming. Whatever customers are trying to do with generative AI foundation models (FMs)—running them, building them, customizing them—they need performant, cost-effective infrastructure that is purpose-built for ML. By leveraging AWS, automotive companies can tap into a choice of chips, models and services with proven security and privacy, and use generative AI to fuel innovation in their operations and the driving experience.
Smart Industry: What unique opportunities are there with generative AI in the auto space?
Wendy: Autonomous driving requires complex software and hardware systems that must be designed to work seamlessly together. Generative AI can be an important tool in designing and testing these systems. For example, generative AI may be used by OEMs to create simulations that test the vehicle’s response to various driving scenarios. These scenarios and the accompanied simulated test data can be edge cases that statistically happen so rarely they are not represented in typical circumstances, or so extreme as to be unsafe to test in the real world (e.g. near miss of a pedestrian crossing at night). This is not just an efficiency improvement, but will also allow automotive companies to create more test scenarios with the potential to improve the overall system capabilities.
Smart Industry: What are you most optimistic about on this front?
Wendy: I’m optimistic about AWS’ approach to creating a space where it's easy for automotive customers to build with generative AI. We’re proud to provide the best tools to support our customers as they build and scale their generative AI-based applications.
