CIO Insider

CIOInsider India Magazine

Separator

EU's AI Rules Demand Industry to be Practice Data Transparency

CIO Insider Team | Friday, 14 June, 2024
Separator

One of the most closely-kept secrets in the industry will have to be revealed as a result of new legislation governing the use of artificial intelligence (AI) in the European Union.

These laws will require businesses to be more transparent about the data used to train their systems.

Roughly 18 months after ChatGPT was first made available to the public by Microsoft-backed OpenAI, interest in and funding for generative AI—a collection of tools that enable the quick creation of text, image, and audio content—have increased dramatically.

However, as the sector grows, concerns have been expressed about how AI firms get the information needed to train their models and whether giving them access to blockbuster novels and Hollywood films without the authors' consent violates their copyright.

Over the next two years, the EU's recently passed AI Act will be implemented gradually to provide regulators time to put the new legislation into effect while businesses adjust to a new set of responsibilities.

A number of well-known tech firms, like Google, OpenAI, and Stability AI, have been sued in the last year by content producers who say their work was inappropriately utilized to train models.

However, it's still unclear how some of these rules will actually operate in practice.

These laws will require businesses to be more transparent about the data used to train their systems.

One of the Act's more controversial provisions is that companies using ChatGPT and other general-purpose AI models give "detailed summaries" of the training material. After consulting with stakeholders, the recently formed AI Office stated that it intends to publish a template for enterprises to use in early 2025.

Even while the specifics are still being worked out, AI businesses are fiercely reluctant to share the data that their models were trained on, characterizing it as a trade secret that, if disclosed, would unfairly offer competitors an advantage.

The level of detail these transparency reports take will have a significant impact on both larger tech companies like Google and Meta, which have made AI the center of their operations, and smaller AI startups.

Current Issue
Defending The Future Indigenously