General AI
AI Tool Claude Aids in Discovering Major Ticketing Flaw
WIRED AIhigh confidence
Why it matters
- →AI tools can autonomously identify and exploit security vulnerabilities.
- →The incident demonstrates the need for robust security audits in digital platforms.
- →AI-assisted security research can lead to improved safety measures.
Security researcher Ian Carroll discovered a critical vulnerability in Front Gate Tickets with the help of AI tool Claude Opus 4.7. This flaw allowed unauthorized access to issue tickets for major US music festivals. Carroll responsibly reported the issue, leading to a quick patch by Front Gate. The incident highlights the potential of AI in identifying security vulnerabilities and the need for robust security measures.
Read originalMore from WIRED AI
© WIRED AIMarket & Regulationbusiness
Apple Sues OpenAI for Alleged Trade Secret Theft
Apple has taken legal action against OpenAI, accusing it of misappropriating trade secrets, including unreleased hardware designs and confidential documents. The lawsuit targets Tang Tan, OpenAI's hardware chief and a former Apple executive, alleging he facilitated the transfer of proprietary technology from Apple to OpenAI. This case could become a landmark intellectual property dispute in Silicon Valley, echoing the Waymo-Uber conflict. The legal proceedings reflect the escalating rivalry between Apple and OpenAI, who were once partners but are now competing in the AI-powered consumer device market.
WIRED AI
© WIRED AIVideo & Creative AIimage
Dataland Opens as First AI Art Museum
Dataland, co-founded by artist Refik Anadol, has opened in Los Angeles as the world's first museum dedicated to AI art. The gallery's debut exhibit, Machine Dreams: Rainforest, uses AI to create immersive experiences that respond to visitors' movements and biometric data. Anadol's work challenges the perception of AI art by emphasizing ethical data sourcing and environmental responsibility. This innovative approach aims to redefine AI art, offering a sensory experience that feels alive and interactive, moving beyond the typical generative art stereotypes.
WIRED AI
© WIRED AIMarket & Regulationbusiness
Anthropic Introduces Usage-Based Fees for Claude Fable 5
Anthropic is shifting its pricing model for Claude Fable 5, moving from flat-rate subscriptions to usage-based fees. This marks a significant change in how consumers access AI models, aligning more closely with developer API billing practices. Subscribers will now pay additional fees based on token usage, potentially increasing costs for heavy users. This move reflects a broader industry trend towards usage-based billing, as AI models become more computationally intensive. While Anthropic aims to return to subscription plans when capacity allows, this change tests consumer willingness to pay for premium AI access.
WIRED AI
More in General AI
© Skill Leap AIGeneral AIwriting
ChatGPT Work Enhances Productivity with New Models
ChatGPT Work is transforming how tasks are accomplished with the introduction of new models like Sol, Terra, and Luna. These models enable users to efficiently create reports, spreadsheets, presentations, and even web apps from a single prompt. The guide demonstrates practical applications such as data cleaning, dashboard creation, and marketing planning, showcasing the versatility of ChatGPT Work. By integrating with desktop apps and utilizing files and folders, ChatGPT Work becomes a powerful tool for business, education, and daily tasks, making AI more practical and accessible for real-world applications.
Skill Leap AI
General AIwriting
Anthropic Introduces Claude Usage Reflection Tool
Anthropic has launched a beta feature that allows users to reflect on their interactions with Claude, their AI assistant. This tool provides a dashboard summarizing usage patterns, helping users understand how AI fits into their daily routines. It encourages users to evaluate when AI is beneficial and when tasks are better handled personally. By offering insights into usage habits and prompting reflection, this feature aims to enhance users' AI fluency and decision-making. This development marks a step towards more mindful and effective AI integration in everyday life.
Anthropic
© TechCrunch AIGeneral AIimage
Google's SynthID Debunks McConnell Deepfake
Google's SynthID system has effectively identified a deepfake image of Senator Mitch McConnell, marking a significant achievement for anti-deepfake technology. The image, falsely portraying McConnell in a hospital bed, gained traction online before Snopes confirmed its inauthenticity using SynthID's invisible watermark. This watermark, embedded in images from participating tools, remains detectable even after being shared widely. The system's success underscores its potential in combating misinformation, though its use is limited to images created with tools that have opted into the program, such as Google's Gemini models and OpenAI's tools.
TechCrunch AI