The EU's AI Act, which is expected to be adopted in April 2024, has received criticism for its shortcomings in adequately protecting human rights. AlgorithmWatch, an organization that analyzes the impact of algorithms on society, has highlighted several missed opportunities in the final law. One of the key concerns raised is the lack of accessibility requirements for low and medium-risk AI systems, which could potentially exclude certain groups from benefiting from AI technologies. Additionally, there are loopholes in transparency obligations for the private sector and security agencies, which could undermine accountability and the protection of individuals' rights. The fundamental rights impact assessments are also deemed insufficient, as they do not provide robust safeguards against potential human rights violations. Another area of concern is the weak remedies for individuals affected by high-risk AI systems, which may limit their ability to seek redress. Furthermore, the AI Act fails to ensure that EU-based AI providers are subject to the same requirements as those inside the EU, potentially creating an uneven playing field. The exemption for national security, partial bans on harmful AI practices, and limited rules on AI use at borders and in migration contexts are also seen as problematic. AlgorithmWatch calls for civil society involvement in the implementation of the AI Act and urges lawmakers around the world to learn from the EU's mistakes.
In a related development, the European Parliament recently approved the Artificial Intelligence Act, which aims to protect fundamental rights and establish Europe as a frontrunner in AI. The law bans certain AI applications that threaten citizens' rights, including biometric categorization systems and untargeted scraping of facial images. It also forbids emotion recognition in the office and educational institutions, social scoring, predictive policing based solely on profiling, and AI that manipulates human behavior or exploits vulnerabilities. Deepfake videos need to be clearly labeled, and Italian PM Giorgia Meloni sought damages after deepfake videos of her were circulated online. Despite the EU's efforts, activists and free speech advocates believe there is room for improvement in the regulations. India is also set to come up with its first draft of AI regulations by June-July. The challenge for India, like in the first world, will be to harness the potential of AI while introducing guardrails that do not stifle innovation.
The European AI Act will be enforced at the end of 2024. The Act classifies AI systems by risk levels. High-risk systems will need to follow transparency requirements. Companies will be responsible for keeping records of the training data. This will force companies to respect data privacy and improve traceability. [49ffb46a] [1af4c27c] [61264d36]