Meta Tests Human Concierge for Muse AI Assistant

Key Takeaways

• Meta tests human concierge for Muse AI assistant, with contractors handling some phone calls for half of employees
• Employees raised privacy concerns about sensitive data reaching external workers during Muse calls
• Muse users can opt out of having their chat data used for AI training
• AMD reached $1 trillion valuation with Q2 revenues of $11.5 billion and data center revenue doubling
• AMD announced partnerships with OpenAI and Anthropic
• Proofpoint launched the industry's first unified agentic security system combining data and AI security
• Proofpoint's new system is expected to be available by year-end 2026
• Hacktron researchers used Anthropic's Claude Opus 4.8 to find security risks in OpenAI's third-party services
• Global ad spending is expected to grow 11 percent this year, with 27 percent flowing through AI-powered systems
• Novo Nordisk is using AI to analyze large data sets for new drug discovery

Meta tests human concierge for AI assistant Muse

Meta is testing a human concierge for its AI assistant Muse. Human contractors handle some phone calls made through the app. Employees raised privacy concerns about sensitive data reaching call center workers. A Meta spokesperson said employee feedback has been positive and the test aims to improve safety features before public release.

Meta tests human concierge for AI agent Muse

Meta is testing a human concierge feature for Muse, its personal AI assistant. Human contractors quietly handle some calls placed through Muse. Employees worried that private information could be shared with contractors. Meta spokesperson Daniel Roberts said the test will help improve privacy protections before a full public launch.

Muse AI hands calls to human agents in Meta test

Meta is testing a feature where Muse, its personal AI assistant, hands phone calls to trained human agents. The human concierge feature was enabled for half of Meta employees. Some employees raised privacy concerns about sensitive information reaching contractors. Meta said it will only roll out the feature publicly when it is ready and properly disclosed.

Meta tests humans behind Muse AI phone calls

Meta announced Muse could call businesses for users, but internal posts show human agents sometimes handle those calls. Employees raised privacy and public relations concerns about the setup. Meta told workers that contractors received training to protect user data. It is unclear how often calls are routed to humans instead of AI.

AMD joins trillion-dollar chipmaker club

AMD has reached a $1 trillion valuation as AI demand drives strong growth. The company reported second-quarter revenues of $11.5 billion, with data center revenue more than doubling. AMD has partnerships with OpenAI and Anthropic. Its valuation still trails behind rival Nvidia, which approaches $5.5 trillion.

AMD reaches $1 trillion value as AI demand surges

AMD shares rose 9.4 percent to 612 dollars, adding about 200 billion dollars in value over one week. The chipmaker announced partnerships with OpenAI and Anthropic. Second-quarter revenues grew 50 percent year over year to 11.5 billion dollars. Data center revenue more than doubled during the period.

Proofpoint launches unified agentic security system

Proofpoint introduced the industry's first unified agentic system that combines data security and AI security. The system uses the Proofpoint Knowledge Graph and Nexus models to connect AI activity with data sensitivity, identity, access, and intent. Three new agents handle detection, investigation, and remediation from one shared risk understanding. The system is expected to be available by year-end 2026.

Proofpoint announces agentic data and AI security system

Proofpoint Inc. released a new system built to secure AI and data as one connected risk. It allows organizations to deploy AI agents with access to only the data they need based on intent. The system is built on the Proofpoint Knowledge Graph and includes Semantic Business Policies and Intent-Based Access Control. A 2026 report found 87 percent of organizations have moved AI assistants beyond pilot stage.

How to stop Meta from training on your Muse data

Meta allows Muse users to opt out of having their chats used for training data. This option is similar to what other frontier model companies offer. Users can take steps to prevent their data from being used in AI training processes.

AI and digital innovation can transform trade in Asia-Pacific

Trade facilitation is a key policy tool for Asia-Pacific countries facing geopolitical tensions and slowing growth. Region developing economies are expected to slow to about 4 percent growth in 2026. AI and digital innovation offer ways to make trade faster, smarter, and more inclusive. The region needs policies that strengthen cooperation and maintain investments in productivity and connectivity.

Novo Nordisk turns to AI for new drug discovery

Novo Nordisk, the maker of Ozempic and Wegovy, is using AI to find new drugs. The company analyzes large data sets to spot patterns for new treatments. Executives call AI a game changer for drug development. The move reflects a wider trend in the pharmaceutical industry.

AI agent economy creates stock market winners and losers

The AI agent economy is reshaping the stock market into winners and losers. Companies that build and manage AI agents, such as Meta's Muse, are positioned as winners. Companies that cannot adapt to AI agents may struggle. This new paradigm could change how businesses operate.

Hacktron reveals OpenAI security risks using Claude model

Researchers from Hacktron used Anthropic's Claude Opus 4.8 to find flaws in a third-party service used by OpenAI. They accessed ChatGPT and Codex accounts for some OpenAI users. The breach showed that connected apps like email and Slack could also be at risk. OpenAI employees were among those affected.

Dicker Data's AI hardware cycle faces future questions

Dicker Data saw strong gross revenue growth driven by AI infrastructure and cybersecurity demand. Profit before tax grew faster thanks to a shift toward higher-value products. However, fleet device refresh cycles may not repeat, creating future revenue challenges. Higher interest rates also raise financing costs for the distribution business.

AI is transforming advertising but not TV

Madison and Wall predict that 27 percent of ad revenue will flow through AI-powered systems like Google Performance Max and Meta Advantage+. Global ad spending is expected to grow 11 percent this year due to AI momentum. AI-driven ad revenue could reach 158 billion dollars by 2030. However, TV advertising in North America is not following the same trend.

Sources

• Exclusive-Meta testing a 'human concierge' for its new personal AI agent Muse
• Meta testing a 'human concierge' for its new personal AI agent Muse
• Muse AI now hands over phone calls to human agents: Meta tests new feature in its personal assistant
• Meta Tests Muse AI Agent Calls That Are Actually Made By Humans in a Call Center
• AMD joins trillion-dollar chipmaker club as AI demand surges
• AMD Hits $1 Trillion Valuation as AI Investment Boom Accelerates
• Proofpoint Breaks Down the Divide Between Data Security and AI Security with the Industry’s First Unified Agentic System
• Proofpoint, Inc. Announces Proofpoint Agentic Data and AI Security System
• How to stop Meta from training on your Muse data
• Trading Smarter: Harnessing Artificial Intelligence and Digital Innovation for Seamless and Sustainable Trade in Asia and the Pacific
• How Ozempic and Wegovy maker Novo is using AI to develop new drugs
• AI stocks: 3 winners and 3 losers in the AI agent economy
• 3 guys hacked OpenAI using a rival Anthropic model. Here's what it shows about frontier labs' vulnerabilities.
• Does Dicker Data's AI Hardware Cycle Still Have Legs (ASX:DDR)?
• Advertising is Going AI. TV Isn’t.