Anthropic sues music publishers while investigating Mythos breach

Key Takeaways

• Anthropic is suing Universal Music Group, Concord, and ABKCO over copyright claims regarding song lyrics used to train Claude, arguing fair use.
• A music publisher representative called Anthropic's use of lyrics blatant infringement and plans to file an opposition brief.
• Anthropic is investigating unauthorized access to its Mythos AI model by a group of AI enthusiasts who gained entry via Discord.
• Anthropic warned that the Mythos model is powerful enough to enable dangerous cyberattacks and is urging users to report suspicious activity.
• Hyperscale Data is expanding its Michigan data center to 83 acres, planning to increase power usage to over 3,000 megawatts and create 500 jobs.
• Representative Seth Magaziner urged proactive federal AI regulation, citing the Mythos breach as a key vulnerability.
• Perpetuals launched BayesShield Clinical to analyze physician diagnostic accuracy using trading methods to reduce the 795,000 annual US deaths from diagnostic errors.
• AI tools are improving ocean cleanup efficiency by 30% by scanning satellite imagery and enabling autonomous vessels to collect plastic 24/7.
• 74% of Americans believe college students should learn AI usage, but a majority oppose using AI as tutors or for screening student applications.
• Toyota unveiled the AI Vision Engine and Integrated ANZEN System for Woven City to improve road safety and coordinate societal impact.

Anthropic fights music publishers over AI training

Anthropic is asking a California federal court to rule in its favor during a copyright lawsuit brought by Universal Music Group, Concord, and ABKCO. The company argues that using song lyrics to train its Claude chatbot constitutes fair use. Anthropic claims this training helps Claude understand human language to improve productivity in science, business, and education. The lawsuit addresses the broader question of whether copying millions of copyrighted works without permission is legally permissible for AI model training.

Anthropic claims Claude AI training is fair use

Anthropic argues that its Claude AI chatbot makes transformative use of song lyrics to understand human language and concepts. The company states that the resulting model can be used for coding, research, and writing in ways unrelated to music. Anthropic asserts that publishers have no evidence showing Claude harms them in the market, which is a key factor in fair use analysis. In response, a music publisher representative called the lyrics usage blatant infringement and promised to file an opposition brief.

Anthropic investigates unauthorized Mythos AI access

Anthropic is investigating an incident where unauthorized users accessed its powerful Mythos AI model. The company has limited the release of the tool due to concerns about potential hacking capabilities and is taking steps to address the security issue. The investigation is ongoing, and Anthropic has not yet disclosed the extent of the breach or specific measures being taken. The company emphasized the need for responsible AI development and will provide updates as the investigation progresses.

Unauthorized users breach Anthropic Mythos AI

A group of AI enthusiasts gained unauthorized access to Anthropic's Mythos AI model via a Discord server. This incident occurred on the same day Anthropic announced a limited release for the model. The individuals were able to test the model's capabilities, including generating websites and executing tasks. Anthropic has warned that the model is powerful enough to potentially enable dangerous cyberattacks, raising concerns about its safety. The company is investigating the breach and urging users to report any suspicious activity.

Hyperscale Data expands AI data center in Michigan

Hyperscale Data is expanding its data center campus in Dowagiac, Michigan, from 34.5 acres to 83 acres. The project includes renovating part of the facility into a combined AI data center and robotics hub with over 100,000 square feet dedicated to robotics assembly and high-performance computing. The site currently uses about 30 megawatts of power but plans to grow to more than 3,000 megawatts. The expansion aims to create more than 500 jobs and combine computing with real-world data collection for AI training.

AI tool detects endangered orcas to reduce noise

OrcaHello is an AI-driven real-time alert system designed to detect southern resident orcas and reduce human disturbance. The tool listens to underwater audio from hydrophones in Puget Sound to identify orca calls and movements. Experts review alerts before confirming them and sending notifications to stakeholders to minimize noise from construction or boat traffic. The system has successfully helped reschedule noisy work and plan fisheries management to protect the endangered subspecies.

Perpetuals applies trading AI to healthcare diagnostics

Perpetuals is launching BayesShield Clinical, an AI tool that analyzes physician diagnostic accuracy using methods originally developed for trading. The system identifies performance patterns to help with case routing, second opinions, and training programs without replacing doctor judgment. The technology targets diagnostic errors, which cause an estimated 795,000 deaths annually in the United States. Perpetuals partnered with the European Institute of Management to manage pilot deployments and trial designs.

Rep Magaziner calls for proactive AI regulation

Representative Seth Magaziner of Rhode Island urged the government to adopt proactive federal policies to govern artificial intelligence. He highlighted the need to ensure AI safety before market release, maintain U.S. competitive advantage, and protect workers from job displacement. Magaziner cited the unauthorized access to Anthropic's Mythos model as an example of potential vulnerabilities requiring regulation. He emphasized balancing AI's potential to improve society with the risks it poses to national security and civil liberties.

AI helps clean oceans more efficiently

Artificial intelligence is helping to clean oceans by scanning satellite imagery to locate plastic and debris patches. AI algorithms transform the cleanup process from guessing into a precise operation, improving collection efficiency by 30%. Autonomous vessels like the Interceptor Original use AI to collect plastic 24/7. These technologies assist with detection and smart removal to make ocean cleanup faster and more scalable.

Survey shows strong support for AI education

Seventy-four percent of Americans believe it is important for college students to learn how to use AI. However, opinions differ on whether students will use the technology to avoid learning or to help them learn. A majority of Americans oppose using AI as tutors or for screening student applications. Despite these views, universities are still struggling to determine appropriate AI policies for the classroom.

Oklahoma City symposium discusses data center impact

An Earth Day symposium in Oklahoma City discussed the environmental impact of data centers and AI. The event coincided with the City Council's approval of a moratorium on new data center projects through the end of the year. Organizers aimed to present facts and ease concerns about energy efficiency and water management. The symposium addressed myths about data centers and emphasized the need for sustainable practices to create a more environmentally friendly society.

Toyota unveils AI technologies for Woven City

Toyota and Woven by Toyota unveiled new AI technologies to drive Kakezan, a strategy for creating greater societal impact at Woven City. The AI Vision Engine uses visual and environmental data to detect risks and enable coordinated action. The Integrated ANZEN System combines this with behavior AI and drive assist to improve road safety. Additionally, the Inventor Garage began operations as a new hub for end-to-end product development and prototyping.

Sources

• Anthropic seeks pivotal court win in music publisher lawsuit over AI training
• Anthropic Claude AI Music Lawsuit: Company Argues for Fair Use Finding
• Anthropic investigating unauthorised access of powerful Mythos AI model
• Anthropic's Mythos AI Accessed by Unauthorized Users
• Hyperscale Data announces expansion plans at Dowagiac data center campus
• AI tool listens for endangered orcas in real time to reduce human disturbance
• Perpetuals extends trading AI to healthcare diagnostics
• Rep. Magaziner: AI Regulation is Needed
• The Transformative Power of AI: Turning the Tide on Ocean Cleanup
• 74 percent say college students should be taught how to use AI: Survey
• Oklahoma City symposium explores data centers' impact on Earth Day
• Toyota and Woven by Toyota Unveil New AI Technologies to Drive Kakezan