This article addresses the gaps and obstacles in the legal and human rights issues surrounding machine learning (AI), as well as how these issues have an impact on basic human rights concepts. These problems include: algorithmic openness, cybersecurity weaknesses, unfairness, bias, and prejudice, lack of contestability, problems with legal personhood, problems with intellectual property, negative effects on workers, problems with privacy and data protection, liability for harm, and lack of responsibility. The article uses the concept of 'vulnerability' to help readers understand significant areas of concern and to direct risk and impact mitigation actions to safeguard people's well-being. This article advances the discussion, which is crucial, while appreciating the good work done in the field of AI legislation and the fact that it requires continual examination and agility in approach.
