Meta issues patch for Muse AI app following zero-day exploit discovery

Published:

Meta has released a patch for its Muse macOS application following the identification of a zero-day vulnerability that could potentially allow unauthorized users to take control of the AI agent. This flaw was discovered by security researcher Patrick Wardle, who noted that it exploited an undocumented setting within Muse, enabling attackers to redirect transcription processing from Meta’s servers to their own systems, as reported by The Verge.

The vulnerability stemmed from several design choices, including the decision to process Muse dictation in the cloud rather than on the device itself. Additionally, the app allowed any application to manipulate all of Muse’s undocumented settings. Wardle demonstrated proof-of-concept attacks that enabled him to take pictures and write malicious files to disk without alerting the user in many instances.

Wardle expressed concerns about the security measures in place, suggesting that the design of Muse could allow attackers to leverage the AI assistant’s privileges for malicious purposes. He indicated that instead of developing complex malware, attackers could simply exploit the AI assistant itself. This criticism contrasts sharply with Meta’s earlier emphasis on Muse’s privacy and security features during its announcement.

In response to the vulnerability, Meta acted swiftly, issuing a patch shortly after the report was published. The company maintains that the real-world security risks were minimal, as the exploit required local access to the user’s device. David Singleton from Meta Superintelligence Labs clarified that this was a local privilege escalation attack, meaning that it necessitated malicious code already running on the user’s machine, thus limiting the practical risk to users of the Muse Mac app.

Despite the rapid resolution of the issue, the exploit has emerged at a challenging time for Meta, as the company seeks to regain its competitive edge in the AI sector. Recently, Amazon blocked Muse from accessing its e-commerce platform, claiming that Meta had not secured permission to do so. Nevertheless, the launch of Muse has seen significant success, with estimated downloads of the mobile app reportedly surpassing those of ChatGPT during its initial 12 days in the US and Canada. Following the launch, Meta’s stock experienced an 11 percent increase.

Follow our Tech news coverage for related developments.

Share post:

Subscribe

Popular

More like this
Related

OpenAI reaches $1.25 trillion valuation after $11 billion funding round

OpenAI has achieved a valuation of $1.25 trillion following a recent funding round that secured an additional $11 billion, as reported by Bloomberg. This significant increase positions OpenAI among the world's most valuable companies, building on its previous valuation of…

North Korea’s elite hackers reportedly net $6.3 billion in Binance heist

North Korea's cyber operations have reportedly netted $6.3 billion from the recent Binance hack, highlighting the regime's sophisticated approach to cybercrime. This development underscores the strategic importance of cyber warfare to North Korea, which has been cultivating a cadre of…

Rolls-Royce highlights advanced turbine blade technology in aircraft engines

Rolls-Royce has unveiled its advanced turbine blade technology, which plays a crucial role in enhancing the performance and durability of aircraft engines. These turbine blades, capable of withstanding extreme mechanical stresses and temperatures exceeding 1,500°C, are produced through a unique…

AT&T and AST SpaceMobile achieve significant milestone in direct-to-device services

AT&T and AST SpaceMobile have reported a significant advancement in direct-to-device (D2D) services, successfully executing a call handoff from a terrestrial network to a satellite network. This milestone, which involved equipment from Ericsson and Nokia, marks a notable step toward…