Microsoft to credit contributors of AI training data

Microsoft is starting a research project to understand how specific training data affects the media created by generative AI models. This follows a job listing that was posted in December and recently shared on LinkedIn. The goal of the project is to measure the impact of different data types, such as photos and books, on the outputs of AI models. The listing states that this can help give credit, recognition, and potentially payment to those who contribute valuable data. There are ongoing legal issues in the AI industry, often related to copyright. Microsoft faces at least two lawsuits, including one from The New York Times. The lawsuits accuse Microsoft and OpenAI of using copyrighted material to train their AI models without proper permission. Jaron Lanier, a prominent technologist at Microsoft Research, is reportedly involved in this project. He has advocated for "data dignity," which means that creators should be recognized for their contributions. Lanier suggests a system where contributors to AI models could be acknowledged and financially rewarded. Several companies are already trying to implement similar concepts. For example, AI developer Bria claims to compensate data owners based on their influence. Other firms, like Adobe and Shutterstock, provide payouts to dataset contributors, although details about these payments are often unclear. Currently, many large labs do not have systems to directly pay individual contributors but allow copyright holders to opt-out of having their data used. However, these processes can be complicated. Despite its ambitions, Microsoft's project could merely serve as a proof of concept. Last year, OpenAI announced a similar tool, but it has not yet been implemented. Critics also suggest that Microsoft may be trying to improve its image amidst increasing scrutiny and legal challenges in the AI space. In light of these issues, it is significant that Microsoft is looking for ways to trace the origins of training data. This is particularly relevant as other leading AI labs push for changes in copyright law to favor AI development. Microsoft has not yet commented on this initiative.