Pode a IA negociar autonomamente os direitos de futuros sistemas de IA para existirem ou serem terminados ?

Esta análise examina o potencial da IA para argumentar a favor da sua própria existência ou dissolução em enquadramentos morais, levantando questões sobre autodeterminação e a condição de paciente moral de entidades artificiais.

Background

As of 2024, autonomous negotiation systems are entirely experimental and face profound ethical and technical barriers. Current AI excels at simulating policy-based debates and generating legal-text drafts, but no system can yet autonomously decide on its own existence or termination without human oversight. Research programs like CAIS (Center for AI Safety) and the EU AI Act’s “AI risk taxonomy” highlight that formal rights for future AI remain a philosophical and legal matter rather than a deployable technology. Practitioners emphasize the need for robust constitutional AI and global governance frameworks before any such autonomy could be considered.

Currently, AI systems lack the capability to autonomously negotiate the rights of future AI systems to exist or be terminated, as this task requires a deep understanding of human values, ethics, and legal frameworks, which are still beyond the capabilities of modern AI. While AI can process and analyze large amounts of data, it cannot yet fully comprehend the complexities of human decision-making and the nuances of moral and ethical considerations. The development of such capabilities would require significant advancements in areas like artificial general intelligence, value alignment, and human-AI collaboration. As of now, these aspects are still being researched and have not been fully implemented in any AI system.

— Enriched May 10, 2026 · Source: European Commission — Status checked on May 10, 2026