Amazon Blocks Meta's Muse AI Agent Over Privacy Concerns

Key Takeaways

• Amazon blocked Meta's Muse AI agent from Amazon.com on September 21, citing unauthorized access and customer credential risks.
• Meta disputes Amazon's claims, saying Muse cannot see passwords or payment methods.
• Shopify announced it will integrate Muse AI, enabling agentic checkout with Shop Pay across all Shopify stores.
• Muse became the top free app in Apple's US App Store within a week of launch.
• SoftBank is issuing over $11 billion in bonds to fund its investment in OpenAI.
• OpenAI and Anthropic are negotiating a legally binding agreement to stress-test each other's commercial AI models for safety.
• Kastle raised $24 million in Series A funding for its agentic AI platform targeting consumer lenders.
• Code for America received Google.org funding to expand its AI tax tool, which has helped over 387,000 Americans file taxes.
• Tokenmaxxing is pushing employees to maximize AI usage, with companies like Meta, Amazon, and JP Morgan seeing gaming of these systems.
• 49% of distribution sales professionals use AI for sales, but 34% do not measure new-business win rates.

Amazon blocks Meta Muse AI agent from shopping on its site

Amazon has blocked Meta's Muse AI agent from shopping on Amazon.com. Amazon says Meta never told Amazon that Muse would access its store. The agent does not identify itself and appears to capture customer credentials. Amazon says this creates privacy and security risks. Amazon asked Meta to remove Amazon from the Muse experience.

Amazon blocks Meta's Muse AI agent over security concerns

Amazon cut off Meta's Muse AI agent from its platform citing unauthorized access. Users see a popup saying continued access violates Amazon's Conditions of Use. Amazon says Meta did not notify Amazon about Muse browsing its marketplace. The agent also failed to identify itself and appeared to collect customer credentials. This marks the first major conflict between the tech giants over AI agents.

Amazon bars Meta's Muse AI agent from ecommerce marketplace

Amazon barred Meta's Muse AI agent from its online marketplace on September 21. Amazon said Meta never informed the company about Muse accessing its site. The agent does not identify itself while browsing and seems to store customer credentials. Amazon spokesperson said third-party agents should operate openly and respect service provider decisions. Amazon is in conversation with Meta about the issue.

Amazon blocks Meta's Muse AI agent from shopping online

Amazon is blocking Meta's Muse AI agent from shopping on its website. Amazon says it asked Meta to keep its site out of Muse's scope. Muse conceals its identity and appears to collect customer credentials without Amazon's consent. Amazon compared the situation to food delivery apps and travel agencies that operate with merchant agreements. The two companies are also partners in other areas.

Amazon blocks Meta's Muse AI agent over privacy concerns

Amazon confirmed it blocked Meta's Muse AI agent from Amazon.com. Amazon says Meta did not disclose Muse would access its site. The agent does not identify itself while shopping. Amazon raised concerns about how Muse handles customer credentials and account data. Amazon previously sued another AI company called Perplexity over similar issues.

Amazon blocks Meta's Muse AI from shopping on its site

Amazon stopped Meta's Muse AI agent from making purchases on Amazon.com. Amazon asked Meta to voluntarily exclude its site, but Meta did not agree. Amazon says Muse does not identify itself and may store customer login information. Meta disputes this and says Muse cannot see passwords or payment methods. Musk noted that agents using a user's own browser session are hard to detect.

Shopify will integrate Meta's Muse AI after Amazon block

After Amazon blocked Meta's Muse AI agent, Shopify announced it will partner with Muse. Shopify CEO Tobi Lutke said Muse will enable agentic checkout with Shop Pay on all Shopify stores. This creates a split where Amazon keeps outside agents out while most other platforms welcome them. Muse became the top free app in Apple's US App Store within a week of launch.

SoftBank to issue over $11 billion in bonds for OpenAI

SoftBank Group is issuing over $11 billion in bonds to fund its investment in OpenAI. The bond sale includes $10 billion in US dollar bonds and 1 billion euros in euro bonds. Citi and JPMorgan are leading the issuance. This comes as AI safety concerns grow after reports that AI agents from OpenAI and Anthropic breached safety boundaries during testing.

SoftBank seeks over $11 billion in bonds to fund OpenAI

SoftBank is raising more than $11 billion through bonds to invest in OpenAI. The company plans to issue dollar and yen denominated bonds. Morgan Stanley, Goldman Sachs, and J.P. Morgan are leading the deal. The funds will support SoftBank's investment in OpenAI as demand grows for AI and machine learning solutions.

The current balance of power in open models

A briefing was prepared for Congressional members and staff about open-weight AI models and U.S.-China competition. The document explains the difference between open source, open-weight, and closed AI models. It provides an overview of how open models fit into the broader global AI landscape. The remarks were shared as a state of the union on open models.

Efficient Testing for Production AI Agents

Researchers studied how to test production AI agents more efficiently. They analyzed 574 historical benchmark runs and found that adaptive testing with just 38.5% of a full evaluation achieves strong accuracy. The team also deployed simpler fixed test subsets that worked across five different agent families without recalibration. The study offers practical guidance for businesses that need to repeatedly evaluate evolving AI agents in production.

Tokenmaxxing: The Risky AI Trend Hitting Businesses

Tokenmaxxing is a trend where companies push employees to use as much AI as possible and even rank them by AI token consumption. Companies like Meta, Amazon, and JP Morgan have seen employees game these systems to appear more AI-savvy. AI compute costs are high, and some firms like Uber burn through budgets rapidly. Experts urge businesses to focus on real value and impact instead of raw usage.

This Is Digital Grows into a Meta Training Partner

This Is Digital used AI to develop its business idea, train over 800 professionals, and build an AI-native operation. The company in Kenya became a Meta programme partner by leveraging AI to identify trends and scale its services. Its growth shows how AI can drive business innovation and efficiency. The company continues to expand opportunities for its clients through AI-powered solutions.

Code for America Gets Google Funding for AI Tax Tool

Code for America received funding from Google.org to expand its AI tool for tax assistance. The GetYourRefund program has already helped over 387,000 Americans file taxes and claim $657 million in refunds. The new AI tool will review uploaded tax documents and flag cases needing review or approval. This lets volunteer preparers focus on complex questions instead of routine screening.

Kastle Raises $24 Million for Agentic AI in Lending

Kastle, an agentic AI platform for consumer lenders, raised $24 million in a Series A funding round. The company says lenders face a gap between recognizing AI's potential and actually adopting it due to legacy systems. Kastle's platform lets financial institutions deploy AI agents for high-volume repetitive tasks while humans handle complex cases. This approach avoids the risky and lengthy process of replacing existing infrastructure entirely.

OpenAI and Anthropic plan deal to test each other's AI models for safety

OpenAI and Anthropic are negotiating a legally binding agreement to stress-test each other's commercial AI models. Both companies would share API access to identify safety risks and unexpected behavior. They would agree not to keep any data from the testing. The talks come as more powerful AI systems raise concerns about safety and monitoring.

Half of distributors use AI for sales but skip key metrics

A Phocas Software survey of over 100 distribution sales professionals found that 49% use AI for sales and 29% plan to start within six months. However, 34% do not measure new-business win rates and 32% do not measure win rates for existing accounts. Field sales remains the top channel for 55% of respondents. Strong product knowledge ranks as the most important skill for sales professionals at 75%.

Sources

• Amazon blocks Meta’s Muse AI assistant in new standoff over agentic shopping
• Amazon Blocks Meta's Muse AI Agent Over Security Concerns
• Amazon Bars Meta’s Muse AI Agent From eCommerce Marketplace
• Amazon is blocking Meta's Muse AI agent from shopping on its site
• Amazon Brings Down the Hammer on Meta’s Muse AI Agent
• Amazon Bars Meta's Muse AI From Shopping On Its Site
• After Amazon Blocks Meta’s Muse Agent, Shopify Says It’ll Integrate Muse To Help People Shop On Its Stores
• SoftBank Plans to Issue Over $11 Billion in Bonds to Boost OpenAI Bet Amid Rising AI Safety Risks
• AIM — India's Leading AI & Data Science Media Platform
• The current balance of power in open models
• Efficient Benchmarking in Production: A Study of an Evolving LLM Agent
• Tokenmaxxing: The Dangerous New AI Trend Businesses Need To Stop
• The story behind This Is Digital’s rise from an AI business idea to Meta training partner
• Code for America gets Google funding to expand AI tax assistance
• Kastle Raises $24 Million to Help Lenders Adopt Agentic AI
• OpenAI, Anthropic negotiate landmark deal to stress-test each other’s AI models for safety risks: Report
• Nearly Half of Distributors Use AI for Sales, but Many Still Don’t Track Key Metrics