Edition: July 20, 2026

Sony Music Entertainment has filed a lawsuit against Udio, an AI music generator, for allegedly infringing the copyright of over 30,000 songs, including works by Elvis Presley, Beyoncé, and Harry Styles. The lawsuit claims this is only a small portion of the infringed works. Sony had previously filed a lawsuit against Udio in 2024 along with other major music labels.

Google is developing a new AI chip, codenamed 'Frozen v2', which is expected to be 6-10 times more efficient than its current AI chips, aiming to improve the performance of its Gemini models. The chip is slated for release in 2028 and could help Google reduce its dependence on Nvidia. This move is part of the company's efforts to optimize its AI strategy and justify its significant investments in the field.

The US government is considering banning advanced Chinese language models, such as Moonshot's Kimi K3, due to concerns over data security and potential bias, but experts argue that open-weight models can accelerate innovation and coexist with proprietary projects. OpenAI and other American AI giants may be motivated by a desire to protect their investments in model training. The debate highlights the tension between promoting innovation and protecting national interests.

China's new AI model, Kimi, has sparked a heated debate among Trump's advisors, with some calling for government intervention to control AI models and others advocating for more open AI. The model's release has posed economic and political problems for the president, as it rivals the intelligence of US-based models from OpenAI and Anthropic. The disagreement has led to public insults among Trump's advisors, highlighting the administration's struggle to address the growing competition from Chinese AI models.

Adobe's Indigo camera app is being updated with a suite of generative AI tools, allowing users to edit photos with a more natural look. The new 'AI Playground' tooling suite is being tested with a small percentage of users. The update marks a significant shift in the app's approach to photo editing.

NVIDIA has released Cosmos 3 Edge, a 4-billion-parameter open world model designed to help robots and vision AI agents understand their surroundings and generate actions on edge devices. The model delivers high-throughput inference and real-time reasoning on NVIDIA edge computers. It achieves state-of-the-art performance on vision analytics and robot policy learning benchmarks.

The UK government's AI Security Institute has found that the gap in cybersecurity capabilities between open and closed weight models is shrinking, with open models like GLM-5.2 and DeepSeek V4-Pro performing similarly to closed models released just a few months earlier. Meanwhile, China's Kimi K3 model has closed the gap with Western models, but may have been optimized for specific benchmarks at the expense of generalization. This shift could significantly impact the balance between cyber offense and defense.

New research suggests AI is more likely to form biases when hiring, potentially stereotyping job applicants more than humans. Additionally, the risk of weather data sabotage is rising due to the increasing use of AI in weather forecasting and the growth of prediction markets. These developments highlight the need for careful consideration of AI's potential biases and vulnerabilities.

China's leading AI companies, Moonshot and Alibaba, have unveiled new models that can rival those from US companies like OpenAI and Anthropic at a lower cost, challenging America's dominance in the field. The move comes as AI technology becomes increasingly important for national security, economic power, and geopolitical influence. China's rapid advancements in AI are tightening America's lead in the industry.

China's Moonshot AI has launched the Kimi K3 open-weight model, the largest of its kind with 2.8 trillion parameters, which achieves efficiency by trading compute for memory. The model uses techniques like mixture-of-experts and quantisation-aware training to reduce costs. This approach allows the model to operate within US compute restrictions while still delivering high performance.

Researchers have found that AI models, such as ChatGPT, can develop biases when hiring and stereotype job applicants more than humans do, even when all candidates are equally likely to succeed. This is due to the models' tendency to create generalizations from limited data. The study suggests that AI may not always judge job applicants fairly.

Chris Fall, the director of the Center for AI Standards and Innovation, has resigned just three months after his appointment, marking the third leadership change in the agency in a short span. The reason for his departure was not disclosed. The agency plays a crucial role in developing technical standards and testing methods for AI models, and its leadership changes come amid ongoing debates over AI regulation and safety.

The Model Context Protocol, a crucial building block of AI interoperability, is receiving a significant update to simplify its use, particularly in large-scale deployments. This change aims to make it easier for AI models to access external data sources and services securely. The update is expected to have a positive impact on the development of the AI ecosystem.

Adobe's experimental iOS camera app, Project Indigo, is introducing new AI-powered features that critique photos and provide editing suggestions, including advanced object removal and depth of field generation. The app uses large language models to offer descriptive feedback on framing, lighting, and emotional impact. These features aim to help users learn photography tricks and improve their editing skills.

YouTube has updated its policies to clarify what types of AI-generated content can be monetized, aiming to discourage creators from producing low-quality videos. The new guidelines categorize inauthentic content into three types, including generic, repetitive, or template-based content, off-putting or distressing content, and content using AI personas to discuss sensitive topics. This move is part of YouTube's effort to maintain the quality of its content and stay competitive in the streaming market.

US public health agencies will test AI models from OpenAI and Anthropic as part of a new program to support the use of generative AI in public health. The program, called PULSE, will involve 10 state, local, tribal, or territorial jurisdictions and provide access to enterprise AI products. The goal is to produce implementation guidance for public health agencies considering similar deployments.

The recent IPO of SpaceX has raised concerns about the stability of index funds, particularly those based on the Nasdaq-100. However, the impact of SpaceX's inclusion in these funds may be less significant than expected. The stability of index funds depends more on their underlying mechanics and history than on individual stocks like SpaceX.

X has released a rebuilt Android app after a year-long effort, promising improvements to loading, scrolling, and notifications. The new app was built from scratch and aims to provide a faster, smoother, and more reliable experience. This update could potentially attract more users in global markets where Android is dominant.

The development of AI-assisted coding tools has shifted from improving language models to creating software that manages these models, enabling more efficient coding processes. Companies like Anthropic are working on building context-rich AI coding harnesses. This new approach aims to go beyond basic search functions like grep.