This shift in stance is attributed to the power of corporate lobbying. OpenAI, for instance, has been lobbying for the watering down of the EU's AI act to reduce the regulatory burden on the company. The EU's decision on whether to allow AI companies to self-regulate is eagerly awaited, with the author expressing hope that the EU will stand firm on its initial proposal.
Key takeaways:
- The EU's AI proposals, which aim to regulate artificial intelligence based on its capacity to cause harm, are entering the final stages of the lawmaking process. However, there is internal disagreement about some key aspects of the proposed legislation, particularly those related to the regulation of "foundation" AI models.
- Foundation AI models, or "general-purpose AI" (GPAI) systems, are important as they form the base for the next phase of the tech future. Any flaws or vulnerabilities in these models could potentially affect the entire networked world.
- There is resistance from big tech companies to the idea of regulators being able to inspect their AI models. Recently, the French, German, and Italian governments have advocated for less intrusive regulation of foundation models, suggesting a regulatory framework that fosters innovation and competition instead of legal regulation.
- The change in stance from these governments is believed to be due to the power of corporate lobbying. It was revealed that OpenAI had lobbied for elements of the EU's AI act to be watered down to reduce the regulatory burden on the company.