In its Opinion printed on 13 October 2022, the EDPS welcomes the opening of negotiations for a Council of Europe conference on synthetic intelligence, human rights, democracy and the rule of regulation (Conference). The EDPS considers the Conference as an necessary alternative to enrich the European Fee’s proposed Synthetic Intelligence Act by strengthening the safety of people’ elementary rights, such because the rights to privateness and to the safety of non-public knowledge.
Wojciech Wiewiórowski, EDPS, stated: “The Conference is a chance to develop the primary legally binding worldwide instrument on synthetic intelligence based on EU requirements and values on human rights, democracy and the rule of regulation. To attain this, the Conference ought to embody applicable, robust and clear safeguards to guard people who could also be affected by way of AI techniques”.
To this finish, the EPDS makes the next key suggestions on the EU’s negotiating directives for the Conference.
The final targets ought to prioritise the safeguards and rights to be supplied to people, and teams of people, which may be impacted by way of AI techniques. To make sure the safety of people and their rights, the negotiating directives ought to goal for consistency of the Conference with the EU’s current authorized framework on knowledge safety.
The EDPS is of the opinion that AI techniques that pose unacceptable dangers to people ought to be prohibited. Extra particularly, the usage of AI for: the social scoring of people; biometric identification of people in publically accessible areas; the categorisation of people on the premise of their biometric knowledge (e.g. their ethnicity); the categorisation of people based on their perceived feelings, ought to be prohibited by default. AI techniques that have an effect on people’ rights to human dignity typically also needs to be prohibited. These situations also needs to be explicitly included within the negotiating directives.
Monitoring the usage of AI techniques, how these techniques work, the danger they entail, is essential. The EDPS recommends that the negotiating directives goal at offering minimal procedural safeguards to guard people which may be affected by way of AI techniques. These safeguards ought to make sure that the usage of AI techniques is clear, could be clearly defined to oversight authorities and people, and could be audited usually to restrict dangers that such instruments could current. Compliance and management mechanisms also needs to be put in place, in addition to cross-border cooperation between competent authorities designed by the events to the Conference, to oversee the usage of AI, and shield people and their elementary rights.
Background data
The principles for knowledge safety within the EU establishments, in addition to the duties of the European Knowledge Safety Supervisor (EDPS), are set out in Regulation (EU) 2018/1725.
The EDPS is the impartial supervisory authority with accountability for monitoring the processing of non-public knowledge by the EU establishments and our bodies, advising on insurance policies and laws that have an effect on privateness and cooperating with comparable authorities to make sure constant knowledge safety. Our mission can be to lift consciousness on dangers and shield folks’s rights and freedoms when their private knowledge is processed.
Wojciech Wiewiórowski (EDPS), was appointed by a joint resolution of the European Parliament and the Council on to serve a five-year time period, starting on 6 December 2019.
The legislative session powers of the EDPS are laid down in Article 42 of Regulation (EU) 2018/1725, which obliges the European Fee to seek the advice of the EDPS on all legislative proposals and worldwide agreements which may have an effect on the processing of non-public knowledge. Such an obligation additionally applies to draft implementing and delegated acts. The statutory deadline for issuing an EDPS opinion is 8 weeks.
Concerning the Synthetic Intelligence Act: On 21 April 2021, the European Fee introduced its Proposal for a Regulation of the European Parliament and of the Council laying down harmonised guidelines on synthetic intelligence. The EDPS and the European Knowledge Safety Board (EDPB) issued a Joint Opinion on this Proposal, which could be discovered right here.
