December brought the ROAD Rome Advanced District team to Milan, specifically to the Plenitude auditorium, for a day of work dedicated to the ‘Data & AI Lab’. After an initial live meeting and numerous remote updates, some forty participants gathered to work as a team.

Around the tables, in addition to the ROAD district members, there were also high-calibre external companies, such as Italgas and Amplifon, who gave the projects a boost with their vision and experience.

Data model and governance: why focus on data?

One of the focuses of the day was the work on a framework designed to answer a crucial question: ‘Why invest in Data Governance?’. The answer came in the form of a cost-benefit analysis model, built to demonstrate in a tangible way how effective data management can turn into business value.

Here is what emerged:

- The benefits: increased efficiency in processes (think cross-selling or marketing campaigns), improved service delivery and a positive impact on customers and users. Furthermore, proper data management can help reduce the risk of penalties related to regulatory non-compliance and lower the costs of access to structured and organised data.

- The costs: data production and management, storage, maintenance and governance costs.

The working group also emphasised that a clear and strategic representation of data is crucial for feeding Artificial Intelligence platforms. In addition, sharing experiences between companies of different sizes helped to identify best practices, reduce application gaps and, why not, learn from the mistakes of others.

Sustainable AI: understanding consumption

The other hot topic of the event was the sustainability of Artificial Intelligence. While AI is an increasingly central tool for companies, its environmental impact cannot be ignored.

The group worked on identifying KPIs to measure and monitor AI-related CO2 emissions. After analysing the existing literature, open source tools capable of measuring the impact of AI training and inference activities were identified.

Three levels of contextualisation were identified:

1.         Infrastructure: energy required to run data centres.

2.         AI use: consumption and solutions to reduce the impact of the use of tools and technologies.

3.         Indirect effects: changes in consumer behaviour with devices that can alter greenhouse gas emissions over time.

The next challenge? To test these open source trackers in the field, analysing both hardware and software components in different application scenarios.

Future roadmap: next steps

The event was not limited to one day of work: it was the starting point to confirm the spirit of an active and dynamic community. Through active involvement, participants were able to make a personal contribution to the overall discussion context. The goal for these working groups is clear: to validate and update the models in 2025.

With its ongoing projects, ROAD continues to prove to be a catalyst for innovation, fostering collaboration between different companies and sectors. Working together is not only a necessity: it is a winning way to meet the challenges of the future.