It has been revealed that OpenAI is offering news organizations between 1 million and 5 million dollars a year to train advanced language models such as GPT-4.
As the world of artificial intelligence (AI) continues to develop rapidly, the value of the data used to train the models is also increasing. This increase in data value is especially evident when it comes to copyrighted content such as news articles.
OpenAI offers news organizations $1-5 million
For comparison, it is known that the figures Meta offers to news organizations for the Facebook News tab are up to 3 million dollars a year. In fact, we recently reported that Google had agreed to pay Canadian publishers to link to their articles.
Concerns about copyright infringement have also led to OpenAI and Microsoft’s models, such as ChatGPT and Copilot, being sued for copyright infringement.
To avoid such problems, AI companies have started to form partnerships with news organizations. Publishers like Axel Springer and Associated Press have signed agreements with OpenAI to train models like GPT-4 and develop news aggregation technology.
OpenAI also has agreements with stock media libraries, such as Shutterstock. However, there is no information about the financial aspects of these agreements.
The cost of training AI models and concerns about copyright are important factors for the future of AI. The debate on this issue must continue, and fair solutions must be found.
Google has not lagged behind developments in this area. The company has introduced leading newspapers such as the New York Times, Wall Street Journal, and Washington Post to a Genesis AI tool that can take facts and generate news stories.
Some news organizations have also started using generative AI tools in their newsrooms. However, there is still debate about the ethics of using these tools and the risk of copyright infringement.
As a result, the value of the data used to train AI models and the risk of copyright infringement create a new tension between AI companies and news organizations. Time will tell how developments in this area will progress and how copyright issues will be resolved.
Featured image credit: Mariia Shalabaieva / Unsplash