Artificial intelligence (AI) and machine learning (ML) are starting to transform how solutions are built and deployed in transportation. In fact, transportation is arguably yielding some of the most ambitious, far reaching enterprise applications of AI and ML. Fully automated transportation technologies such as autonomous vehicles and self-driving sea vessels capture the public’s imagination; but these technologies require further development and will likely involve safety regulation before widespread adoption. More quietly, however, AI and ML is being used extensively in the transportation industry, especially to streamline supply chain operations (for example, in AI-enhanced real-time supply chain planning). There are also a range of applications of AI and ML being used in fleet management such as for route optimization, vehicle supply management, and promoting safety.
The applications of AI and ML can be very valuable. They can be leveraged to protect market share and also to create new revenue streams through commercialization. However, achieving success in these activities requires that the companies making these investments ensure that they own or control the technologies or associated IP rights, and that a range of risks – including a number of important ones centred on legal and IP issues – are managed effectively.
AI/ML is generally deployed in transportation collaboratively because a variety of technology or data inputs from different parties are required. Such collaboration requires collaborative development agreements, which can be complex to negotiate and require care, especially around the impact on ownership or rights to the use of technology and data. If multiple projects and parties are likely to be involved, there may be tremendous benefit in designing and implementing a technology/data consortium with associated template agreements and policies that enable projects to proceed in an agile fashion, while managing IP and data related issues responsibly.
AI/ML technology for transportation generally requires a wide range of private and public data sources to develop and train the technology. A key issue in regards to data is what is being increasingly called “data provenance”: where is data coming from, and what are the terms of use that apply and their implications. The best practice is usually to undertake “data due diligence,” similar to that undertaken for any other asset, which is both a technical and legal analysis directed to addressing the risk ramifications of use and reliance on specific data sources.
AI/ML technology often involves the use of open source technology or data. Commercialization can also involve the use of open licensing techniques. This requires resolution of the collective implications of the various applicable rights regimes regarding ownership or rights to use key assets. Increasingly, organizations are using a set of tools and processes to navigate recurring issues such as mapping tools to reassess implications of open source resources on a project as it evolves.
Protecting the value created through investments in AI/ML can benefit from IP protection. But IP for IP’s sake may not always be productive. Also, particularly in AI/ML, the optimal mix of IP tools deployed, whether patents or trade secrets for example, can be technically challenging to set. Usually a bespoke mix of tactics – possibly including an integrated licensing strategy – works best.
Especially in transport, where there can be substantial human risks if systems fail, a robust risk-based approach to mapping and mitigating risks is critical, and legal, IP, data and regulatory risks contribute in a substantial way to applicable risk profiles. Increasingly, enterprises are learning that ensuring that AI-enabled solutions comply with evolving legal and ethical standards is not just important to manage risks, but also to ensure commercial success. Acceptance and adoption by customers necessitates a robust approach to legal/ethical compliance. Failure to comply can result in backlash and even injury to the broader business reputation.
Therefore, a key take away is the need for a bespoke protection and commercialization strategy for AI/ML innovations in transportation; one that takes a risk-based approach and accounts for the broad range of legal, regulatory, ethical, IP and data considerations.
