London’s Financial Times has entered into a “strategic partnership and licensing agreement” with OpenAI, the company behind the AI chatbot ChatGPT. This agreement allows OpenAI to access and use the FT’s content for training its AI models and generating responses through tools like ChatGPT. While the financial terms of the deal remain undisclosed, it is clear that the FT sees value in collaborating with OpenAI to explore the potential of generative AI and develop new AI products and features for its readers.
The FT’s CEO, John Ridding, emphasizes the significance of this partnership, stating that it recognizes the value of the newspaper’s journalism and provides insights into how AI can surface content. He also highlights the importance of transparency, attribution, and compensation for publishers when it comes to AI platforms using their material. This sentiment aligns with OpenAI’s aim to address legal liabilities around copyright by paying publishers for their copyrighted content.
OpenAI’s move to license content for model training with publishers like the FT demonstrates its effort to address the issue of “hallucination” in large language models (LLMs), which can generate false or inaccurate information. By leveraging reputable journalism from publishers, OpenAI hopes to improve the accuracy and reliability of its models.
Aside from mitigating legal risks and improving model accuracy, publishers also stand to gain financially from these licensing deals. OpenAI has already signed several publisher agreements and has more in the pipeline. While monetizing their content is beneficial, news publishers must also consider the potential impact of generative AI on their traffic and readership. As users of ChatGPT may click on citations that link to publishers’ content, it could drive traffic to their sites. However, if generative AI becomes a substitute for search engines, it could divert traffic away from news publishers.
Publishers must exercise caution in their involvement with AI technologies to avoid reputational pitfalls. CNET, for example, faced backlash when it adopted generative AI for content production without clearly disclosing its use to readers. The FT, known for its quality journalism, will need to carefully integrate generative AI into its products and newsroom processes to maintain its reputation.
In Europe, the use of AI tools like ChatGPT is further complicated by legal uncertainties surrounding privacy laws. Publishers and AI companies must navigate these concerns to ensure compliance and protect user privacy.
Overall, OpenAI’s partnership with the Financial Times highlights the growing interest and potential of generative AI in the publishing industry. By collaborating with reputable publishers, OpenAI aims to enhance the accuracy of its models while providing publishers with compensation for their content. As this technology continues to evolve, it will be crucial for publishers to strike a balance between leveraging AI’s capabilities and safeguarding their reputation and readership.