TL;DR:
- Major newspapers are in talks with OpenAI for access to digital news stories.
- Publishers demand a share of the $1.3 trillion generative AI market.
- Over 535 news organizations block content for ChatGPT training.
- Discussions revolve around compensating publishers for content access.
- Reddit and Elon Musk also seek payment for their data.
- Generative AI leads to copyright lawsuits and collective bargaining efforts.
- Shutterstock sets an example with a Contributor Fund.
- Advocacy for copyright protections gains momentum.
Main AI News:
A wave of negotiations is underway between prominent newspapers and OpenAI, the pioneering creator of ChatGPT, concerning access to a crucial resource in the age of generative artificial intelligence: digital news stories. Traditionally, technology firms like OpenAI have freely harnessed news articles to build datasets that instruct their machines in comprehending and engaging with human queries. However, as the race to develop cutting-edge AI models intensifies, newspaper publishers and other data owners are now asserting their share of the potentially lucrative generative AI market, predicted to reach a staggering $1.3 trillion by 2032, as per Bloomberg Intelligence.
Since August, over 535 news organizations, including industry giants such as The New York Times, Reuters, and The Washington Post, have implemented blockers to prevent the collection and utilization of their content for ChatGPT’s training. The current focus of discussions centers around compensating publishers, enabling the chatbot to incorporate links to individual news stories within its responses. This dual-pronged approach offers direct financial benefits to newspapers and the potential for increased web traffic.
In a strategic move, OpenAI secured a content licensing deal with the Associated Press to leverage its data for AI model training. The ongoing discussions have touched upon this concept but have predominantly emphasized integrating news stories into ChatGPT’s responses.
Beyond newspapers, other valuable data sources are seeking compensation. Reddit, a popular social message board, has engaged with leading generative AI companies regarding payment for its data. In a daring move, Reddit contemplates blocking search crawlers from Google and Bing, which could hinder the forum’s discoverability but is deemed worthwhile in exchange for compensation.
Notably, in April, entrepreneur Elon Musk instituted a $42,000 fee for bulk access to Twitter posts, a service previously available to researchers for free, citing illegal data usage by AI companies. This reflects a growing sense of urgency and ambiguity surrounding the monetization of online information. With generative AI poised to revolutionize internet interactions, many content providers view fair compensation as a fundamental concern.
For instance, shortly after the launch of GPT-4 by OpenAI in March, the coding community Stack Overflow witnessed a 15% reduction in traffic as programmers increasingly relied on AI for coding solutions, leading to layoffs. Additionally, prominent AI firms are confronting copyright lawsuits from authors, artists, and coders seeking damages and a portion of the profits.
OpenAI’s decision to engage in negotiations may stem from the desire to establish agreements preemptively, avoiding potential legal disputes over content licensing and payment obligations for tech companies. According to James Grimmelmann, a digital and information law professor at Cornell University, OpenAI maintains that its practices have not infringed upon copyright law and is primarily interested in future content access beyond fair use.
The generative AI sector has witnessed a substantial influx of venture capital, with nearly $16 billion invested in the first three quarters of 2023. This massive funding underscores the substantial costs associated with developing AI technology, encompassing hardware, computing power, and, until recently, freely accessible data.
As OpenAI and Google implemented tools to block their AI data crawlers, online platforms like Reddit, Stack Overflow, and Wikipedia initiated protective measures. They now offer paid access to their data for AI training and impose limits on data mining frequency.
While individual departures may not sway AI operators, publishers are seeking strength in numbers. Industry leaders are exploring collaborative efforts to secure compensation from AI companies, recognizing the importance of asserting their intellectual property rights.
In this evolving landscape, Shutterstock, a stock photo platform, exemplifies an innovative approach. It has partnered with OpenAI to provide training data and established a Contributor Fund to remunerate artists. Although the fund has distributed over $4 million, the median payout per image remains relatively modest.
Danielle Coffey, President and CEO of the News/Media Alliance (NMA), representing over 2,000 publishers, highlights the receptive stance of the White House and policymakers regarding the necessity for licensing deals. She has undertaken advocacy efforts to secure copyright protections for publishers, emphasizing the integral role of quality content and journalism in the realm of generative AI.
Conclusion:
The negotiations between newspapers and OpenAI, along with other data owners’ demands for compensation, signal a pivotal shift in the market. The generative AI landscape is evolving towards a more balanced ecosystem, where content providers assert their rights and seek fair compensation, potentially reshaping the dynamics of the industry.
