Desafíos actuales de la Inteligencia Artificial

46 Desafíos actuales de la Inteligencia Artificial copyright protection, watermarking 76 , and reporting of energy consumption. 77 Providers of generative AI models that pose systemic risks are required to conduct model evaluations, perform risk assessments (including red teaming), fulfill reporting obligations for serious in- cidents, and ensure cybersecurity measures are in place. 78 None of these obligations specif- ically target harms related to discrimination. It is considered that although risk assessments and mitigation efforts should encompass bias and non-discrimination, it remains to be seen whether the harmonized standards (Articles 40 and 41) and the codes of practice (Article 56) will effectively tackle discrimination issues in practice. 79 These standards will determine which types of bias are covered and the remedies that must be implemented. 80 To summa- rize, the essential provisions of the AI Act are aimed at non-generative AI, and regrettably, there are no explicit requirements imposing developers of generative AI models to design systems that mitigate bias. 81 (2) Risk Assessment: A Data Protection Impact Assessment (DPIA) concentrates on how risks related to per- sonal data are alleviated, whereas a Fundamental Rights Impact Assessment (FRIA) has a broader scope, assessing the degree of impact on a variety of fundamental rights such as free- dom of expression, access to justice, and the right to good administration. 82 This distinction is highlighted by the controllers’ obligation under Article 25(1) of the GDPR to implement 76 Article 53(1)(a) and 55, European Parliament legislative resolution of 13 March 2024 on the proposal for a regulation of the European Parliament and of the Council on laying down harmonised rules on Artificial Intelligence (Artificial Intelligence Act) and amending certain Union Legislative Acts (COM(2021)0206 – C9- 0146/2021 – 2021/0106(COD))(Ordinary legislative procedure: first reading).References to articles and recitals introduced or changed by the Parliament will be labelled in this paper as “the EU AI Act”. 77 AnnexXI, SectionII(2)(e) European Parliament legislative resolution of 13 March 2024 on the proposal for a regulation of the European Parliament and of the Council on laying down harmonised rules on Artificial Intelligence (Artificial Intelligence Act) and amending certain Union Legislative Acts (COM(2021)0206 – C9- 0146/2021 – 2021/0106(COD))(Ordinary legislative procedure: first reading).References to articles and recitals introduced or changed by the Parliament will be labelled in this paper as “the EU AI Act”. 78 Article55(a)and(d), European Parliament legislative resolution of 13 March 2024 on the proposal for a regula- tion of the European Parliament and of the Council on laying down harmonised rules on Artificial Intelligence (Artificial Intelligence Act) and amending certain Union Legislative Acts (COM(2021)0206 – C9-0146/2021 – 2021/0106(COD))(Ordinary legislative procedure: first reading).References to articles and recitals introduced or changed by the Parliament will be labelled in this paper as “the EU AI Act”. 79 Hacker, Zuiderveen Borgesius, Mittelstadt, and Wachter, ‘Generative Discrimination’, p. 40. 80 Hacker, Zuiderveen Borgesius, Mittelstadt, and Wachter, ‘Generative Discrimination’, p. 40. 81 Hacker, Zuiderveen Borgesius, Mittelstadt, and Wachter, ‘Generative Discrimination’, p. 40. 82 A. Mantelero, ‘The Fundamental Rights Impact Assessment (FRIA) in the AI Act: roots, legal obligations and key elements for a model template’ (2024) p. 32; T. E. Loeber Lutz Riede, Christoph Werkmeister, Philipp Roos, Lena Isabell, ‘EU AI Act unpacked #6: Fundamental rights impact assessment’ (June 2024).