OpenAI Accused of Training AI Models on Copyrighted Content

•

a day ago

Summary

The nonprofit organization AI Disclosures Project has accused OpenAI of training its AI models on copyrighted content without permission. The report claims that OpenAI's GPT-4 model was trained on nonpublic books, including works from authors such as J.K. Rowling and Stephen King. This raises concerns about the potential misuse of copyrighted material in AI development.

Key Points

The report accuses OpenAI of training its models on copyrighted content without permission.
The report claims that OpenAI's GPT-4 model was trained on nonpublic books, including works from authors such as J.K. Rowling and Stephen King.
This raises concerns about the potential misuse of copyrighted material in AI development.

Why It Matters

The use of copyrighted content without permission can have significant legal and ethical implications for AI developers, users, and creators.

Author

Kyle Wiggers

More Headlines

Epic Games CEO Calls Out Apple, Google

Epic Games CEO Tim Sweeney called out Apple and Google for their 'gangster-style' businesses and illegal practices during a Y Combinator event. He emphasized how these companies' practices directly affected his own business, scaring away users from installing Epic's Games Store software and preventing Epic from attracting developers to its offerings.

OpenAI Lays Out Its Profit Transition Plans

As OpenAI prepares to transition from a nonprofit to a for-profit, the company is convening an expert group to help its philanthropy understand the most urgent and intractable problems nonprofits face today. The group will provide insights to OpenAI's board of directors within the next 90 days.

TikTok Buyers: Mnuchin, Oracle, Walmart, Microsoft, Rumble, and Perplexity AI

The United States is trying to find a buyer for TikTok, after the Trump administration banned the app's parent company ByteDance. Mnuchin, Oracle, Walmart, Microsoft, Rumble, and Perplexity AI are reportedly interested in acquiring TikTok.

Rippling Publicly Releases Affidavit in Corporate Espionage Case Against Deel

Rippling has publicly released an affidavit detailing allegations of corporate espionage against Deel. The affidavit claims that the highest levels of Deel's leadership are implicated in a scheme to facilitate Russian payments and harass employees who try to blow the whistle.

Meta Combines Instagram Accounts of Second Lady and Former Second Gentleman

Meta is combining the Instagram accounts of Usha Vance, the Second Lady of the United States, and Doug Emhoff, the former Second Gentleman. The 1.2 million followers of @secondgentleman46archive will be transferred to @SLOTUS. This is part of a regular process for presidential transitions.

Circle Files for USDC Stablecoin IPO

Circle, the company behind USDC stablecoin, has filed for an initial public offering (IPO). In its regulatory filing, Circle reported $1.68 billion in revenue and $156 million in net income for 2024. This is the second attempt at listing on the stock exchange by the company.

Roblox Adds New Safety Measures

In a bid to prevent grooming, explicit content and violent material, Roblox has updated its parental controls. The new features allow parents to block access to certain experiences or people on their kid's friends list. They can also track which experiences their child spends the most time in.

Patreon Partners with Sony Music and Amazon Wondery for Podcasting

Patreon, a subscription platform for online creators, has signed podcasting partnerships with Sony Music Entertainment and Amazon's Wondery. The partnership will make some of the podcasting industry's big names available on Patreon, allowing the site to stand out among other platforms that support podcasts.

Parasail Aims to Disrupt AI Infrastructure Landscape

Parasail works with dozens of providers to deliver on-demand GPUs for companies and enterprises looking to build AI models and applications. The founders aim to disrupt the AI infrastructure landscape by providing a more horizontal, fragmented approach.

AI Sales Rep Startups Are Booming, But VCs Are Wary

AI sales representative (SDR) startups are a crowded market these days. While some of these startups are certainly growing fast, the field has its challenges and some VCs are wary.

Anthropic Launches Claude for Education

Anthropic has launched Claude for Education, an AI chatbot designed specifically for higher education. The new tier offers additional capabilities such as 'Learning Mode' to help students develop their critical thinking skills.

Roam's Assumable Mortgage Platform

Roam's platform allows buyers to find over 2,000 assumable mortgages in Houston alone, reducing fall-through rates and increasing acceptance rates. The startup also claims that it can speed up the process of becoming a homeowner, closing deals in just 45 days.

The Fediverse Ramps Up Its Security With New Vulnerability Disclosure Program

The Fediverse, an open social web that includes Mastodon, Threads, and Pixelfed, is ramping up its security. The Nivenly Foundation has launched a new security fund to pay for responsible disclosures of vulnerabilities in the Fediverse. This program aims to improve the security posture of the open social web.

DeepMind Publishes Paper on AI Safety

DeepMind has published a comprehensive paper on AI safety, outlining its approach to mitigating the risks associated with Advanced General Intelligence (AGI). The paper warns of severe harm and existential risks if AGI is developed without appropriate safeguards. DeepMind's approach focuses on blocking bad actors' access to hypothetical AGI, improving understanding of AI systems' actions, and hardening environments in which AI can act.

Redwood Materials Expands to San Francisco

Redwood Materials, founded by former Tesla CTO JB Straubel, is expanding its presence in San Francisco with a new research and development center. The 15,000-square-foot facility will be equipped with lab space to support engineers who will work on every point of the battery ecosystem.