News

30/06/2022

"Sandbox", a virtual environment to implement responsible artificial intelligence

By Helga Zanotti

The European Commission proposal on April 2021 for a regulation on artificial intelligence opened the door to monitoring AI through "sandbox," a controlled virtual environment in which potential risks arising from its use are monitored. The main areas reviewed were health, safety and fundamental rights.

Spain offered itself as the pilot for the "AI sandbox" project aimed at achieving the following objectives:

- clarify and flesh out the requirements outlined in the proposed AI regulation for AI systems;

- support compliance with know-how by applying the AI regulation to companies that are already implementing it or planning to implement it;

- encourage innovation and enable the development of reliable AI systems;

- build expertise and start a path of consultations in Spain for the creation of a national supervisory authority;

- test future obligations and fulfillments arising from the activity in a controlled environment, and drafting guidelines to develope both national and European standards.

The collaboration between Spain and the European Commission will be close and continuos; however, other European countries are also allowed to assist or participate directly. This gives the sandbox a pan-European dimension.

The sandbox will feature two focus groups:

1. the first one will implement in practical terms the digital environment with a focus on the regulatory aspect and testing the different solutions, also collaborating with different authorities;

2. the second one will prepare the overall structure of the sandbox, study its development and share the know-how when implementing the requirements (including technical ones).

Spain will develope the role of scientific hub of excellence on the topic of digital innovation, already developed by the Spanish National AI Strategy. Europe, on the other hand, seems to be limited to the role of liaison between innovators and regulators, for the creation of a controlled digital environment that facilitates development, testing and validation, paying attention to compliance with the requirements that the proposed AI regulation identifies, as well as any national standards that will be promulgated.

For more information, please contact Paula Vega at the following e-mail address: p.vega@bmvinternational.com.




To stay updated on BMV news, publications and initiatives, subscribe to our newsletter