The Cabinet Office’s Central Digital and Data Office (CDDO) recently launched a new algorithmic transparency standard for all government and public services.
Indeed, this national standard aims to meet the commitments made in the national data strategy and the national AI strategy. In doing so, the UK also hopes to strengthen its position as a trusted global leader in AI.
This was put in place after the Center for Data Ethics and Innovation (CDEI) recommended a mandatory transparency requirement for public sector organizations that use algorithms to support important decisions affecting people. .
In addition, this standard has been strongly supported by civil society organizations and international organizations which have always advocated for better transparency. With this, they hope that the risks associated with algorithmic decision-making will be managed and more scrutinized in order to build public confidence.
The standard will then have two levels. The first will include a brief description of the algorithmic tool, as well as how and why it is used, while the second will contain more detailed information on how the tool works, the datasets that were used to train the model. and the level of human oversight.
Thus, the standard aims to help teams to be more transparent about the way in which algorithmic tools are used in order to support decisions, in particular those concerning individuals.
The CDDO will review the standard in the following months before seeking formal approval from the Data Standards Authority in 2022.