Welcome to the age of generative AI, where leading tech conglomerates seem to be preaching one thing but practicing another when it comes to online content usage.

Titans of tech such as OpenAI, backed by Microsoft, and Google, including its partner Anthropic, have been utilizing online content from various companies to train their generative AI models. This procedure hasn’t involved requesting explicit permissions, leading to an impending legal showdown that’s set to shape the future of web and copyright laws in the AI era.

The tech world may claim that this approach is under the purview of ‘fair use,’ but this is still up for debate. Interestingly, while these tech giants readily use others’ content, they prohibit the use of their own data for training rival AI models. A question thus arises: why the double standard?

Looking at Anthropic’s AI assistant, Claude, their terms of service clearly restrict the use of their services for developing competing products or services, including AI or machine learning models. The same goes for Google’s generative AI, and OpenAI’s ChatGPT, both of which prohibit using their services to develop rival machine learning models.

These companies certainly understand the importance of quality content in training effective AI models, hence their strict policies. But this raises another question: why should other companies and websites allow their content to be used freely by these tech giants?

So far, OpenAI, Google, and Anthropic have not responded to requests for comments on this issue.

Several companies are only now realizing this stark discrepancy and they’re less than pleased. Reddit, for instance, has been a significant resource for AI model training over the years and now plans to monetize access to its data.

“We don’t need to give all of that value to some of the largest companies in the world for free,” expressed Steve Huffman, Reddit’s CEO.

In April, Elon Musk fired accusations at Microsoft, OpenAI’s primary backer, for allegedly using Twitter’s data unlawfully to train AI models. Meanwhile, Sam Altman, OpenAI’s CEO, is reportedly working on AI models that respect copyright, suggesting that creators might be paid for the use of their content or style.

Some publishers, including News Corp., are also advocating for tech companies to pay for utilizing their content to train AI models. This interest is unsurprising given that publishers have a significant stake in this matter.

Steven Sinofsky, a former Microsoft executive, expressed concern that the current method of AI training “breaks” the web. He emphasized that the creators or copyright holders receive no value in return for their content used to train AI models.

As we navigate through the complex landscape of AI, questions about ethical data usage and fair compensation are more relevant than ever. Companies, creators, and investors must stay vigilant, ensuring that the journey into the future of AI doesn’t compromise the rights and interests of those whose content fuels these AI models.