Defang AI, or make it toothless?

economictimes.indiatimes.com:

When it comes to AI, Asimov's 'force' can be interpreted as blunt restrictions, or rejection of the technology. But this approach doesn't truly solve the issue at hand, but may lead to more conflicts and challenges. On June 14, the European Parliament approved its position on the proposed AI Act.

This Act, adopted by a large majority, aims to create the first set of concrete regulations on AI, charting the course of its development and deployment within the EU. Its text is poised to guide inter-institutional negotiations with the EU Council and the European Commission. This could play a critical role in shaping the world's approach to AI regulation.

The AI Act's categorisation of risk levels illustrates the kind of balanced regulatory approach that can foster innovation in sectors like healthcare, transportation, energy and manufacturing, while ensuring safeguards against AI's potential hazards. These hazards range from jeopardising physical security and causing economic losses, to raising ethical dilemmas. In this context, it becomes easier to understand why HAL 9000, the AI gone rogue in Stanley Kubrick's 1968 film, 2001: A Space Odyssey, based on the Arthur C   Clarke short story, 'Sentinel', would fall into the category of 'unacceptable risk' — of an AI system prioritising its programmed objectives over human safety — under the AI Act.

The European Parliament has significantly amended the draft of the Act first introduced in 2021, creating a seismic shift in the EU's legislative roadmap to institute a comprehensive legal framework. It widens the classification of high-risk AI systems to include those leveraged to sway voters and recommender systems on very large online platforms (VLOPs). This would help in curbing