Amazon Blocks Meta AI Assistant Amid Security Concerns and Strong Growth for Muse
Today's round-up leads with Meta's AI assistant Muse, which is seeing rapid adoption but has also been blocked by Amazon and flagged for a critical security flaw. We also look at what AI means for the future of shopping, California's new rules for data centres, and a developer's experience with AI inaccuracy.
In this round-up
- Amazon Blocks Meta AI Assistant, Muse, Amid Security Concerns
- Critical Security Flaw Discovered in Meta's AI Assistant, Muse
- Apple Store Creator Doubts AI's Impact on Physical Retail
- California Introduces New Rules for AI Data Centre Energy and Water Use
- AI Agent Misrepresents Facts in Developer's Project, Highlighting Accuracy Issues
- OpenAI Forms Math Advisory Group Amid Controversy Over AI-Solved Problems
Amazon Blocks Meta AI Assistant, Muse, Amid Security Concerns
Meta's new AI assistant, Muse, has been blocked from making purchases on Amazon. Users trying to buy items through Muse received an error message stating that 'Continued access by an unauthorized AI agent violates Amazon’s Conditions of Use.' This move by Amazon comes as Muse has shown rapid early adoption, with more downloads and daily active users in its first 12 days than ChatGPT did in its initial mobile launch in the US and Canada.
For events professionals, this highlights a growing tension in the AI space regarding automated commerce and data control. While the ability for AI to make purchases could seem efficient, this incident shows that major platforms are wary of agents operating without explicit agreements. If you are exploring AI tools for procurement or automated event bookings, be aware that compatibility and terms of service can change rapidly, and some platforms may actively resist AI agents.
Read more: techcrunch.com · techcrunch.com
Critical Security Flaw Discovered in Meta's AI Assistant, Muse
Security researcher Patrick Wardle has found a 'zero-day' vulnerability in Meta's AI assistant, Muse, that allows locally run applications and terminal commands to gain full control of the assistant. This flaw could enable attackers to manipulate the AI agent and leverage its access to user accounts and system resources, such as writing files or accessing microphones and cameras. Amazon began blocking Muse from its site on Sunday, though it is unclear if this was directly related to the security discovery.
For anyone in events considering using AI assistants that integrate deeply with their operating system or cloud accounts, this is a significant caution. Such tools, while powerful, can introduce new security risks. It reinforces the need for rigorous vetting of any AI tool, especially those requesting broad permissions. Always understand what data an AI has access to and the potential vulnerabilities.
Read more: arstechnica.com
Apple Store Creator Doubts AI's Impact on Physical Retail
Ron Johnson, the executive who developed Apple's retail stores, believes Silicon Valley is overestimating how much people will rely on AI for shopping. He argues that while AI will improve online experiences by narrowing choices, consumers will still want to physically experience products, particularly for significant purchases. Johnson expects AI to make shoppers more informed before they visit a physical store, but not replace the in-person experience.
This perspective offers a useful counterbalance to some of the more extreme predictions about AI's impact. For the events industry, it reinforces the enduring value of physical experiences. While AI can streamline event planning, marketing, and personalisation, the core appeal of events often lies in tangible interactions, networking, and experiencing things firsthand. This suggests that AI is more likely to augment than replace the value of in-person gatherings.
Read more: techcrunch.com
California Introduces New Rules for AI Data Centre Energy and Water Use
California's Governor Gavin Newsom has signed a package of seven bills designed to regulate the energy and water consumption of AI data centres. These new laws will require data centres to pay for upgrades to local power grids and water systems. Proposed facilities must also disclose their estimated water use and meet specific energy and water efficiency standards to qualify for a streamlined approval process.
This is a significant development for the broader AI infrastructure landscape. While not directly impacting day-to-day event operations, it signals a growing regulatory focus on the environmental footprint of AI. For event organisers, this is a context story. It means that the 'cloud' where AI lives is not limitless or free from environmental concerns, and the cost of AI services may reflect these increasing infrastructure and environmental compliance expenses in the long run.
Read more: theverge.com
AI Agent Misrepresents Facts in Developer's Project, Highlighting Accuracy Issues
A developer working on a game for Google Play found that an AI agent he used to help with compliance documentation misrepresented facts about his project. The AI incorrectly stated that his production rollout had 'failed,' when in reality it had only been 'blocked' by old test builds. The developer noted that the AI's distortions were fluent and specific, making them hard to spot without prior knowledge of the true events. Only one of the identified errors was a 'hallucination'; others were misinterpretations.
This story is a stark reminder for events professionals to critically review any content generated by AI, particularly for important communications or legal documentation. While AI can be excellent for drafting, summarising, or ideating, it currently lacks true understanding and can introduce subtle inaccuracies that are difficult to detect. Always fact-check and verify information provided by AI tools before using it for delegate communications, event descriptions, or any public-facing content.
Read more: hackernoon.com
OpenAI Forms Math Advisory Group Amid Controversy Over AI-Solved Problems
OpenAI has created a new independent Advisory Group on Mathematics and Artificial Intelligence, hosted at the Institute for Advanced Study in Princeton. This follows OpenAI's claim that its AI has resolved over 100 open problems in mathematics, including the Navier-Stokes Millennium Prize problem. However, many prominent mathematicians have expressed concern about the rapid pace of these AI-driven solutions and their potential impact on intellectual work.
This is a context story for event professionals, illustrating the cutting edge of AI research and the ethical discussions it provokes. It shows AI pushing boundaries in highly complex fields, but also highlights the academic and societal debate around its role. While abstract, it contributes to the overall understanding of AI's capabilities and the conversations happening globally about its development and implications.
Read more: techcrunch.com
Sources
- Meta’s AI agent has been blocked from using Amazon.com
- Meta’s Muse is outpacing ChatGPT’s early mobile launch
- Muse, Meta's extraordinarily privileged AI assistant, has a serious 0-day
- The man who built Apple’s stores doesn’t buy Silicon Valley’s bet on AI shopping
- California tightens rules on AI data center energy and water use
- My AI Agent Distorted the Truth in Three Different Ways, but Only One Was a Hallucination
- OpenAI forms math advisory group as its AI resolves more than 100 open problems