Harnessing AI to safeguard fisheries

Toward more agile decision-making

About one-fifth of the world’s fishing catch is illegal, costing coastal nations billions of dollars each year. However, innovative solutions are on the way. Soon, law enforcement may leverage the power of AI algorithms to identify illegal fishing in European waters. How? By analyzing data on catches of endangered species, information from ship logbooks, and geolocation of vessels. This is just one example of many analytics solutions envisioned by the Directorate-General for Maritime Affairs and Fisheries (DG MARE), the European Commission department that ensures sustainable use of ocean resources, among other responsibilities.

These innovative analytics solutions will be made possible by a department-wide digital transformation underway at DG MARE. Accenture is supporting the transformation by migrating DG MARE’s existing IT applications to Amazon Web Services’ (AWS) cloud platform. As part of the collaboration, Accenture’s Data and AI Studio for Public Service in Europe is helping DG MARE use generative AI to reduce development time. The goal? Equip DG MARE to carry out its mission more quickly and effectively.

Making business rules cloud-compatible

DG MARE’s IT applications handle a wide range of data, such as fishing vessel locations and fish catch declarations submitted by the 27 European Union member states as well as “non-member third party countries.” Historically, applications and their data have been siloed. This challenges DG MARE’s ability to cross-reference and analyze different data sets to support the enforcement of its policies and to provide transparency to EU citizens. The result of the cloud migration will be a single platform, called Ocean Store, that unifies and integrates all the data from the legacy applications.

One of the legacy applications, a business rule engine, manages and executes more than 2,000 business rules. The rules validate that data sent by member states and third countries to DG MARE is correct and complete. For example, the business rule engine validates whether incoming data on fish catches is linked to a specific month. If it’s not, the rule engine tells the country how to correct the data before it can be accepted.

The cloud migration requires rewriting the legacy code for the 2,000 rules so that the rule engine can take full advantage of cloud services. This “refactoring” process is time-intensive and would take human developers an estimated 12,000 person-days. To speed up the process, we developed a generative AI tool that uses large language models to understand the rules’ logic and automatically generate new, cloud-ready code. One unique feature of the tool is that it uses DG MARE’s existing codebase and documentation, enabling high-quality code.

Productivity gains and analytics

Based on our work so far, we’ve found that the tool can refactor each rule in just one minute with 90% accuracy. The developer just needs to validate the code and, in some cases, correct inaccuracies.

The tool’s success sparked the team’s interest in identifying new innovative use cases leveraging both AI capabilities and DG MARE’s integrated data. The productivity gains allow the team to spend more time exploring analytics solutions.

For instance, one potential analytics use case under consideration involves comparing complex navigation patterns and catches reported across fishing vessels and automatically identifying unusual or suspicious behavior.

Accenture’s support of DG MARE’s digital transformation is allowing the department to be more agile in data-driven decision-making.