A new initiative is aiming to bring transparency and accountability to the world of artificial intelligence (AI) by evaluating and certifying AI products for copyright compliance. The nonprofit organization, Fairly Trained, was founded by Ed Newton-Rex, who stepped down from his position at Stability AI due to concerns over AI “exploiting creators.”
Similar to organic certifications for agricultural products or fair trade certifications for cocoa, the Fairly Trained label will be awarded to AI companies that have obtained proper consent for the data used to train their AI systems. This certification process provides customers with the assurance that the AI products they are using are copyright-compliant.
In an industry where the use of copyrighted training data is often a contentious issue, Fairly Trained is seeking to establish a new standard. While some AI leaders argue that the use of data without a license falls under the “fair use” exception to copyright law, many rights holders disagree and have initiated lawsuits against AI companies.
The initiative has already certified nine small startups that have demonstrated their commitment to using licensed training data. By obtaining permission and paying a small submission fee, these companies have gained the Fairly Trained certification, allowing them to use the label on their AI products.
One of the main goals of Fairly Trained is to provide consumers and businesses with the ability to choose AI models that align with their values. Ed Newton-Rex believes that there is a mutually beneficial setup possible between AI companies and rights holders, which can be achieved through proper licensing and transparency.
Advisers to Fairly Trained include industry experts such as Maria Pallante, CEO of the Association of American Publishers, and Tom Gruber, a co-founder of the team that created Siri. Their expertise and guidance will help shape the future of copyright compliance in the AI industry.
By establishing clear guidelines and certifying AI products for copyright compliance, Fairly Trained aims to create an environment where AI companies can thrive while respecting the rights of creators. This initiative marks an important step toward ensuring ethical and responsible use of AI technology.
The source of the article is from the blog maltemoney.com.br