• Written by Sonali Subhadarsini, Product Manager

Some people object to the use of enterprise AI governance in automotive, citing concerns that the complexity of the industry will render such systems ineffective, but this objection is misplaced, as the real challenge lies not in the complexity itself, but rather in the specific ways that automotive supply chains and engineering processes can complicate the implementation of AI governance, particularly when it comes to controls that hold up when the system is used by people who did not build it, under the constraints of tiered supply with tight takt and long programmes.

Unique challenges of automotive

The automotive industry presents several unique challenges that can make implementing effective enterprise AI governance more difficult, including supplier escalations, which can involve multiple tiers of suppliers and complex communication protocols, engineering change propagation, which requires careful management to ensure that changes are properly implemented and documented, and warranty root cause analysis, which can be particularly difficult due to the complexity of modern vehicles and the need to identify the root cause of problems that may have arisen from a variety of sources.

For example, a shift supervisor at a manufacturing plant may need to investigate a quality control issue that has arisen on the production line, which could involve querying the system to identify the source of the problem, and then escalating the issue to a supplier if necessary, all while ensuring that the proper protocols are followed and that the issue is properly documented, which can be a complex and time-consuming process, especially if the system is not properly designed to handle such scenarios.

Enforcing permission at retrieval

One key aspect of effective enterprise AI governance is enforcing permission at the point of retrieval, rather than relying on instructions to the model, which can be difficult to implement and may not always be effective, particularly in a complex industry like automotive, where multiple users may need to access the system and retrieve information, and where the system may need to be able to handle a wide range of queries and requests.

By enforcing permission at retrieval, the system can ensure that users only have access to the information they need to perform their jobs, and that sensitive information is properly protected, which is particularly important in the automotive industry, where intellectual property and trade secrets may be at risk, and where the consequences of a security breach could be severe, as can be seen by visiting our website for more information.

Auditability

Auditability is another critical aspect of enterprise AI governance, as it allows organisations to track and monitor system activity, including what was asked, what was returned, and on what data, which can be essential for ensuring compliance with regulatory requirements and for identifying and addressing potential security risks, particularly in a complex industry like automotive, where the consequences of a security breach could be severe.

For example, the category manager at a parts supplier may need to track and monitor system activity to ensure that the system is being used properly and that sensitive information is not being accessed improperly, which can be a complex and time-consuming process, especially if the system is not properly designed to handle such scenarios, and where the consequences of a security breach could be severe, as can be seen by contacting us at contact-us for more information.

Routing to a human

There are certain categories of queries that should always be routed to a human, rather than being handled by the system, including sensitive or confidential information, complex or nuanced queries that require human judgment, and queries that involve multiple stakeholders or require coordination across different teams or departments, which can be particularly important in the automotive industry, where the consequences of a mistake or miscommunication could be severe.

For example, a query about a potential safety issue with a vehicle may need to be routed to a human, rather than being handled by the system, to ensure that the issue is properly investigated and addressed, and that the necessary steps are taken to protect public safety, which can be a complex and time-consuming process, especially if the system is not properly designed to handle such scenarios.

Limitations of enterprise AI governance

While enterprise AI governance can be a powerful tool for managing and controlling AI systems, it is not a panacea, and there are certain limitations and potential drawbacks that organisations should be aware of, including the potential for the system to be overly restrictive or cumbersome, which can limit its effectiveness and make it more difficult for users to access the information they need, and the potential for the system to be vulnerable to certain types of attacks or exploits, which can compromise its security and integrity.

Organisations should carefully consider these limitations and potential drawbacks when implementing enterprise AI governance, and should work to ensure that the system is properly designed and configured to meet their specific needs and requirements, which can be a complex and time-consuming process, especially in a complex industry like automotive, where the consequences of a mistake or miscommunication could be severe.