OpenAI confirmed that it has entered into a new strategic partnership with the Financial Times, in order to use the portal’s news content to train the company’s AI models.
In a joint announcement, the two entities confirmed this Monday the partnership, which involves providing OpenAI with access to the Financial Times’ journalistic content, which will be used to train the company’s various AI models.
The content will be used within the various AI services that OpenAI provides, namely for ChatGPT. At the same time, when information is used based on content from the Financial Times, an indication of this will appear in the responses provided, also containing the link to the original content.
This agreement involves OpenAI paying a sum to the Financial Times to gain access to the content. Although the final value was not revealed for this agreement, at the beginning of the year some sources indicated that OpenAI would be willing to pay between 1 and 5 million dollars per year to use content from some journalistic sources to train its AI models.
Until now, the main generative AI platforms have used information that is publicly accessible on the internet, but there are some questions about how this content is used to train AI models. At the same time, the platforms that train these models aim to obtain quality sources with factual information to be able to train them with correct data, and news sources are one of the main ways to obtain these details.
It should be remembered that platforms such as The New York Times and the BBC have already accused OpenAI of collecting and using their content without authorization to train existing AI models, with some legal actions on the table between the companies involved.
Bear in mind that the Financial Times provides content accessible through payment systems – paywall – with prices of up to 39 dollars per month. This agreement with OpenAI could, theoretically, allow unlimited access to the platform’s news content without having to pay that amount.
