9 Amendments of Svenja HAHN related to 2020/2016(INI)
Amendment 23 #
Draft opinion
Paragraph 1
Paragraph 1
1. Considers that AI used by police and judicial authorities has to be generally categorised as high-risk, given that the role of these authorities is to defend the public interest; considers that the EU should take the lead in laying down basic rules on the development and use of AI by public institutions to ensure the same high level of consumer protection across the EU;
Amendment 34 #
Draft opinion
Paragraph 2
Paragraph 2
2. Stresses that AI should help to ease the administrative burden on public authorities, without ever fully replacing human decisions, and that AI systems should rely on human oversight;
Amendment 42 #
Draft opinion
Paragraph 3
Paragraph 3
3. Considers that such tools should be released as open source software under the public procurement procedure, and that a fundamental rights audit should be part of a prior conformity assessment; believes that – while ensuring the respect of EU law and values and the applicable data protection rules, and without jeopardising investigations or criminal prosecutions – training data must always be open data;
Amendment 49 #
Draft opinion
Paragraph 4
Paragraph 4
4. Emphasises that data collection and the monitoring of individuals should be limited to criminal suspects and court approved surveillance;
Amendment 51 #
Draft opinion
Paragraph 4 b (new)
Paragraph 4 b (new)
4 b. Highlights that national authorities should receive training and basic skills to deal with algorithmic systems and responsibly use AI technologies in criminal matters, with the aim of protecting European citizens from potential risks and damages to their fundamental rights;
Amendment 62 #
Draft opinion
Paragraph 6
Paragraph 6
6. Recalls the high risk of abuse of certain types of AI, including facial recognition technologies in public spaces, automated behaviour detection and profiling to divide people into risk categories at borders, and calls on the Commission to ban themir use by public institutions;
Amendment 65 #
Draft opinion
Paragraph 7 a (new)
Paragraph 7 a (new)
7 a. Calls for exchanges of information and best practices regarding the application of AI techniques and tools by judicial and police authorities in Member States to avoid a fragmented approach in the Single Market, as well as to face in a coordinated manner the risks associated with AI technologies, such as vulnerability to cybersecurity threats, and ensure the protection of citizens in the Union;
Amendment 67 #
Draft opinion
Paragraph 7 a (new)
Paragraph 7 a (new)
7 a. Emphasises that where decision making is assisted by statistical calculations, such as at probation hearings, the decision makers need to be trained about the general biases statistical calculations carry and made aware about the specific biases of calculation in the particular situation;
Amendment 68 #
Draft opinion
Paragraph 7 b (new)
Paragraph 7 b (new)
7 b. Recalls the right of rectification established in Regulation (EU) 2016/679 (General Data Protection Regulation) and stresses the particular importance of accurate data sets, when these are used to assist administrative decisions; calls on the Commission to examine the benefits of ensuring transparency regarding the individual data included in the particular calculation and an accompanying procedure for rectification.