LIVE · cybersecurity feed
Live wire
Is AI Making Us Worse at Cybersecurity?CVE-2026-114713 · Attackers Exploit UnrealIRCd Flaw That Enables Server-Level Command InjectionCVE-2024-10441 · Synology BeeStation OS Vulnerability Linked To Ransomware CampaignsCVE-2020-1337 · Microsoft Windows TOCTOU Race Condition CVE-2020-1337 Linked to RansomwareCVE-2024-58406 · Huatian Dynamic OA8000 Path Traversal Flaw Exploited Upon DisclosureCVE-2026-124218 · Addonsorg Drag and Drop File Upload for Elementor Forms Flaw Exploited Upon DisclosureCVE-2026-114713 · UnrealIRCd Improper Neutralization Flaw Exploited Upon DisclosureCVE-2026-114713 · U.S. CISA adds Check Point and UnrealIRCd flaws to its Known Exploited Vulnerabilities catalogFBI flash alert says Artex tool automating cyberattacks at ‘machine speed’CVE-2026-84869 · SCKit abuses ScreenConnect to deploy remote access and target technician sessions
ai

Who watches the AI watching your street?

Yusaku Fujii, a professor at Gunma University in Japan, has developed a system designed to audit and penalize the misuse of artificial intelligence that analyzes street camera footage. His proposal, called the Verifiable Record of AI Output (VRAIO), aims to ensure accountability for AI systems operating in what he terms Fully Monitored Public Spaces (FMPS), areas with dense camera coverage…

ZeroDay News ·

Source: Help Net Security

Yusaku Fujii, a professor at Gunma University in Japan, has developed a system designed to audit and penalize the misuse of artificial intelligence that analyzes street camera footage. His proposal, called the Verifiable Record of AI Output (VRAIO), aims to ensure accountability for AI systems operating in what he terms Fully Monitored Public Spaces (FMPS), areas with dense camera coverage capable of continuous tracking of individuals and vehicles.

The VRAIO system introduces an independent third party, referred to as the Recorder, which manages an outbound firewall for municipal AI systems. Before any AI output leaves the system, the operator must declare the cameras used, the time range, and the purpose of the data access. The Recorder then verifies this declaration against predefined legal rules and logs its decision in a tamper-resistant ledger. Crucially, the Recorder does not possess decryption keys and cannot access the content of the footage itself; it only processes the labels and declarations.

A key vulnerability in this design is that the Recorder cannot verify the truthfulness of an operator's declaration. For instance, an operator could claim to be searching for a missing child but use the footage for an entirely different purpose. The Recorder would approve this request if the stated purpose aligns with the rules. To address this, Fujii's system incorporates unannounced spot checks. During an audit, the recorded declaration is compared against the actual data released. If a discrepancy is found, the ledger identifies the request and the operator responsible, who could then face administrative penalties, criminal liability, and public disclosure of the violation. The effectiveness of this deterrent depends on the frequency of audits and the severity of the penalties, which Fujii suggests should be adjusted based on experience to ensure the expected cost of misuse outweighs its potential benefits.

Fujii acknowledges several limitations of the VRAIO system. Bypassing the firewall would leave no record, requiring conventional security measures to detect. The Recorder itself could be compromised by the operators it oversees. Furthermore, the system does not evaluate the fairness of the rules; if a city implements discriminatory rules, VRAIO would enforce them faithfully. Fujii envisions the Recorder as an independent body, possibly a judicial institution or an entity between administrative and judicial systems, overseen by multiple parties to ensure its impartiality.

Beyond technical safeguards, Fujii emphasizes the importance of public acceptance for FMPS. He defines true acceptance as when individuals express approval, attribute it to a lack of harm or a perceived benefit, and exhibit no changes in their routes, information-seeking behavior, expression, or participation in assemblies. He distinguishes this from mere resignation, where individuals accept tracking simply because avoidance is impossible. An earlier experiment involving 11 cameras and notifications to 2,218 households yielded positive opinions from residents, but did not measure behavioral changes, which Fujii now considers essential. He proposes anonymous surveys to gauge behavioral impacts and also suggests measuring positive effects, such as increased feelings of safety and greater use of public spaces, particularly for vulnerable populations.

Fujii's plan for implementing FMPS involves three stages. The first stage focuses on locating missing children and consent-based child safeguarding. The second stage involves using AI in streetlights to detect crimes or accidents. The third, which he calls a "stress test," would involve a central AI learning behavioral patterns from past crimes and flagging pedestrians who resemble these patterns. This final stage would first be tested through vignette studies and simulations, with pre-defined stopping criteria including a substantial rise in "chilling effects," unacceptable false detections, or a significant increase in complaints or withdrawals. Fujii has not yet discussed his proposal with regulators or municipalities, advocating for its development and testing through a small-scale demonstration project.

ai
ShareXLinkedInWhatsAppFacebook

More News

view all →
breach

Cursor confirms security breach led to malicious Claude Code extension

The AI-powered code editor Cursor has confirmed a significant security breach that resulted in the distribution of malicious software through its marketplace. The incident, which occurred on October 29, involved the compromise of credentials belonging to the AI infrastructure provider Fireworks AI. These credentials allowed unauthorized actors to access internal Cursor systems, leading to the…

aicritical

Is AI Making Us Worse at Cybersecurity?

The integration of artificial intelligence into cybersecurity operations has sparked a debate regarding whether the technology is inadvertently degrading the core competencies of security professionals. While AI is widely acknowledged for its ability to accelerate data processing and increase operational efficiency, concerns are mounting that an overreliance on these tools may be eroding the…

ai

The Security Jobs Growing Around Enterprise AI Agents

The rapid integration of AI agents into enterprise workflows has shifted the focus of corporate security from simple chatbot oversight to the complex management of autonomous, high-privilege digital assistants. While AI adoption was previously confined to isolated chat windows, modern assistants now possess the capability to read shared mailboxes, access internal databases, and execute tasks…

ransomware

Beware the New AI-Fueled Ransomware Boom

The landscape of ransomware and data extortion is undergoing a significant shift as cybercriminals increasingly integrate artificial intelligence into their operations. While fully autonomous, AI-driven ransomware has not yet materialized, researchers have observed a clear trend of attackers using large language models to streamline and enhance various stages of the attack chain. This…

security

Upcoming Speaking Engagements

Security researcher Bruce Schneier has announced his upcoming public speaking schedule for the latter half of November 2026. The updates, published on November 14, 2026, outline two specific engagements taking place in the province of Ontario, Canada.

security

US disrupts North Korean campaign behind $400 million crypto theft

The United States government has successfully disrupted a sophisticated, 18-month-long social engineering campaign identified as FaceStealer, which authorities link to a North Korean threat actor responsible for the theft of over $400 million in cryptocurrency. The operation, which federal investigators and private cybersecurity researchers attribute to the group known as BlueNoroff, utilized…