Most artificial intelligence projects don't go into production due to the lack of good-quality data. One of our core technical principles at Clearbox AI is “Good AI needs good data."

In today’s rapidly evolving AI landscape, enterprises are exploring innovative solutions to valorize their information assets to develop robust, fair, and explainable AI models. One such solution is synthetic data.

We are very proud to announce the recent start of a new exploratory collaboration with Poste Italiane, the biggest Italian goods and services distribution network. During the exploration phase, we will create and test a synthetic data sandbox to innovate the corporation's management processes and valorize its informative assets.

Testing synthetic data

After an in-depth analysis of their data needs and strategy, Poste Italiane decided to experiment with synthetic data, which will play a crucial role in solving problems caused by scarce, dirty, or unbalanced data in the coming years. We are enthusiastic about this technology's present and future implications: According to Gartner's analysis, by 2024, 60% of the data used in AI projects will be synthetically generated. Moreover, the European Data Protection Supervisor has dedicated a task force to synthetic data, and the MIT Technology Review included it among the most promising new technologies.

Synthetic data retains the original data's statistical properties and predictive power, making it realistic. It ensures the developed models are consistent and explainable, making their functioning more effective in a real corporate environment. This technological development aligns with the draft AI Act Regulation, currently under approval by EU legislators, where the issues of model biases and interpretability are central and crucial.

The exploratory collaboration with Poste Italiane

The innovative project between Clearbox AI and Poste Italiane aims to create a synthetic data sandbox, i.e. a delimited and controlled environment fueled by synthetically generated data. It can be useful to improve models' training and testing by many Poste Italiane divisions and enable aggregated statistical investigation and exploration of its information assets in a secure environment.

We are very pleased to see that Poste Italiane’s vision of technological development aligns with the recent requirements of the European AI Act. In fact articles 10.4 and 54.1 of the AI Act equate synthetic data to non-personal data for AI model training as well as testing with data sandboxes.

The OPEN ITALY program

Clearbox AI and Poste Italiane connected through the framework of OPEN ITALY, the collaborative innovation program of Elis Innovation Hub.

OPEN ITALY is a project dedicated to enterprises, startups and innovation enablers. The program allows large companies and innovative entities to work together in developing new solutions, fostering an open innovation culture. The program originated in 2017 from a semester project of the ELIS Consortium, a meeting on innovation and development for society.

The collaboration between Clearbox AI and Poste Italiane started during the 6th edition of the program, which involves the participation of 60 major Italian companies ready to innovate through the technological capabilities of startups. Among the reasons that prompted the corporation to initiate this initiative is the possibility of conducting concrete experiments with innovative companies.

The corporation engaged in discussions, analysis, and proposals of its innovation needs within the program. Among these needs, a strong interest emerged in researching and experimenting with new solutions - especially artificial intelligence - that can successfully innovate the management of Poste Italiane's information assets.

A stimulating project

At Clearbox AI, we are excited to start this stimulating project with Poste Italiane, a significant step in applying AI technologies in the Italian corporate system. The opportunity to work with such an established entity allows us to demonstrate the potential of synthetic data in practice, offering innovative and secure solutions for the adoption of robust and transparent models. Moreover, the framework of OPEN ITALY represents an important step in our country's innovation landscape. We look forward to contributing our technological know-how to build the future of data and AI together.