You shall understand what you regulate: A case study of the ANPD-Meta case in Brazil
Resumo
Regulating artificial intelligence is a primary need of current legal systems. However, bridging legal, policy and technical expertises to achieve good regulation is hard. We illustrate this difficulty by doing a short case study of a decision by the Brazilian Autoridade Nacional de Proteção de Dados–ANPD suspending Meta’s collection of data to train its large language models. I try to demonstrate that the lack of technical knowledge led the ANPD to issue a bad decision, that ended up backfiring. I end by discussing implications for regulatory best practices.
Palavras-chave:
Artificial Intelligence Regulation, Data Protection, Brazilian Data Protection Authority (ANPD), Policy and Technology Integration, Regulatory Best Practices
Referências
Autoridade Nacional de Proteção de Dados (2024a). Despacho Decisorio nº 20/2024/PR/ANPD.
Autoridade Nacional de Proteção de Dados (2024b). Despacho Decisorio nº 33/2024/PR/ANPD.
Bond, S. and Allyn, B. (2021). Facebook whistleblower tells Congress products hurt kids and weaken democracy.
Fried, I. (2024). Scoop: Meta won’t offer future multimodal AI models in EU.
Guha, N., Lawrence, C. M., Gailmard, L. A., Rodolfa, K. T., Surani, F., Bommasani, R., Raji, I. D., Cuéllar, M.-F., Honigsberg, C., Liang, P., and Ho, D. E. (2023). The AI Regulatory Alignment Problem. Technical report, Stanford Human-Centered Artificial Intelligence.
Meta (2023). The AI behind unconnected content recommendations on Facebook and Instagram.
Naveed, H., Khan, A. U., Qiu, S., Saqib, M., Anwar, S., Usman, M., Akhtar, N., Barnes, N., and Mian, A. (2023). A Comprehensive Overview of Large Language Models. Journal of Latex.
Ouyang, L., Wu, J., Jiang, X., Almeida, D., Wainwright, C. L., Mishkin, P., Zhang, C., Agarwal, S., Slama, K., Ray, A., Schulman, J., Hilton, J., Kelton, F., Miller, L., Simens, M., Askell, A., Welinder, P., Christiano, P., Leike, J., and Lowe, R. (2022). Training language models to follow instructions with human feedback.
Sennrich, R., Haddow, B., and Birch, A. (2016). Neural Machine Translation of Rare Words with Subword Units. In Proceeding of the 54th Annual Meeting of the Association for Computational Linguistics, pages 1715–1725, Berlin. ACL.
Team Llama @ Meta AI (2024). The Llama 3 Herd of Models. Technical report, Meta.
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, and Polosukhin, I. (2017). Attention is all you need. Advances in Neural Information Processing Systems, 2017-Decem(Nips):5999–6009.
Wei, J., Tay, Y., Bommasani, R., Raffel, C., Zoph, B., Borgeaud, S., Yogatama, D., Bosma, M., Zhou, D., Metzler, D., Chi, E. H., Hashimoto, T., Vinyals, O., Liang, P., Dean, J., and Fedus, W. (2022). Emergent Abilities of Large Language Models. Transactions on Machine Learning Research.
Wimmer, M. (2024). PROCESSO Nº 00261.004509/2024-36, Voto n.º 11/2024/DIRMW/CD.
Autoridade Nacional de Proteção de Dados (2024b). Despacho Decisorio nº 33/2024/PR/ANPD.
Bond, S. and Allyn, B. (2021). Facebook whistleblower tells Congress products hurt kids and weaken democracy.
Fried, I. (2024). Scoop: Meta won’t offer future multimodal AI models in EU.
Guha, N., Lawrence, C. M., Gailmard, L. A., Rodolfa, K. T., Surani, F., Bommasani, R., Raji, I. D., Cuéllar, M.-F., Honigsberg, C., Liang, P., and Ho, D. E. (2023). The AI Regulatory Alignment Problem. Technical report, Stanford Human-Centered Artificial Intelligence.
Meta (2023). The AI behind unconnected content recommendations on Facebook and Instagram.
Naveed, H., Khan, A. U., Qiu, S., Saqib, M., Anwar, S., Usman, M., Akhtar, N., Barnes, N., and Mian, A. (2023). A Comprehensive Overview of Large Language Models. Journal of Latex.
Ouyang, L., Wu, J., Jiang, X., Almeida, D., Wainwright, C. L., Mishkin, P., Zhang, C., Agarwal, S., Slama, K., Ray, A., Schulman, J., Hilton, J., Kelton, F., Miller, L., Simens, M., Askell, A., Welinder, P., Christiano, P., Leike, J., and Lowe, R. (2022). Training language models to follow instructions with human feedback.
Sennrich, R., Haddow, B., and Birch, A. (2016). Neural Machine Translation of Rare Words with Subword Units. In Proceeding of the 54th Annual Meeting of the Association for Computational Linguistics, pages 1715–1725, Berlin. ACL.
Team Llama @ Meta AI (2024). The Llama 3 Herd of Models. Technical report, Meta.
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, and Polosukhin, I. (2017). Attention is all you need. Advances in Neural Information Processing Systems, 2017-Decem(Nips):5999–6009.
Wei, J., Tay, Y., Bommasani, R., Raffel, C., Zoph, B., Borgeaud, S., Yogatama, D., Bosma, M., Zhou, D., Metzler, D., Chi, E. H., Hashimoto, T., Vinyals, O., Liang, P., Dean, J., and Fedus, W. (2022). Emergent Abilities of Large Language Models. Transactions on Machine Learning Research.
Wimmer, M. (2024). PROCESSO Nº 00261.004509/2024-36, Voto n.º 11/2024/DIRMW/CD.
Publicado
27/11/2024
Como Citar
PADUA, Joao Pedro.
You shall understand what you regulate: A case study of the ANPD-Meta case in Brazil. In: CONFERÊNCIA LATINO-AMERICANA DE ÉTICA EM INTELIGÊNCIA ARTIFICIAL, 1. , 2024, Niteroi.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2024
.
p. 137-140.
DOI: https://doi.org/10.5753/laai-ethics.2024.32471.
