Methodology:
The research utilized a mixed-methods approach, examining AI incidents and their impacts on human rights. Data was sourced from the AIAAIC Repository, which documents AI-related incidents. Qualitative analysis was furthered through literature reviews, analysis of human rights instruments, AI regulations, and expert interviews, ensuring a comprehensive grasp of the topic.
Challenges posed by the "AI Abyss", meaning the gap between AI's rapid development and its widespread adoption.
- Rapid Evolution and Marketability: AI's fast growth makes it difficult to regulate, especially concerning ethics, bias, and privacy.
- Human-Like Behavior and Bias: AI systems can exhibit biases, discrimination, and stereotypes similar to humans.
- Data Privacy and Accuracy: AI requires vast data, posing privacy challenges, especially with generative AI that learns from varied sources.
- Multifaceted Impact of AI: AI affects human rights at multiple stages, from data collection to deployment.
Key Findings:
- Increase in AI Incidents: A significant rise in AI-related incidents in the past decade has been observed, particularly affecting vulnerable groups.
- AI's Complex Impact: AI's effects on human rights are multifaceted, arising from technical and societal issues.
- AI's Vulnerable Targets: Incidents majorly affect vulnerable groups, such as children, persons with disabilities, and racial/ethnic minorities.
- From Incidents to Impacts: A direct link is needed between an AI incident and a human rights violation, based on international law standards.
- Need for Inclusive Regulation: While there is some regulation, human rights must play a more dominant role in AI discussions. The research also highlighted an increase in AI-related incidents, especially affecting women, children, persons with disabilities, and racial minorities.
Conclusion:
There is an immediate need to bridge the gap between AI advancements and human rights protection, especially for vulnerable groups. Addressing the challenges of the AI abyss and implementing the provided recommendations can steer us towards an AI future that respects and enhances human rights.
