One open model now sits months behind the American frontier on cyber and biology, and refused nothing it was asked to do. The closed model refused so often the test could not be finished. Months of capability separate them. The gap in restraint is total. Part four of four.
The White House finished its frontier AI framework on 1 August and has published nothing. The threshold is classified. The benchmarks are classified. Whether open weight models are covered at all remains unanswered. Part three of four on governing what cannot be recalled.
Washington is pushing its AI security perimeter deep inside the data centre, targeting Chinese-made components that move data between GPUs. The policy may reduce cyber and espionage risks, but it could also raise costs, slow construction and expose a new weakness in America’s AI race as AI scales.
A zero-day vulnerability in Palo Alto Networks' PAN-OS, CVE-2024-3400, is under active attack, allowing unauthenticated root access. The flaw affects versions 10.2, 11.0, and 11.1, with patches due by April 14, 2024. Users should enable Threat Prevention ID 95187 or disable telemetry.
A zero-day vulnerability in Palo Alto Networks' PAN-OS software, labelled CVE-2024-3400, has been under attack since March 26, 2024, well before it was publicly disclosed.
Users of affected devices should enable Threat Prevention Threat ID 95187 if that is available, otherwise, disable device telemetry until patches are available from the vendor, per vendor instructions. Base Score: 10.0 CRITICAL
This flaw, which received a CVSS severity score of 10.0, allows unauthenticated attackers to run arbitrary commands with root access on affected firewalls.
It affects versions 10.2, 11.0, and 11.1 of the PAN-OS software when configured with GlobalProtect gateway and device telemetry.
This ongoing exploitation, referred to as Operation MidnightEclipse, is being tracked by Palo Alto Networks' Unit 42. It involves a command injection vulnerability that threat actors have used to install a cron job.
This job periodically executes commands from an external server, specifically "172.233.228[.]93/policy" or "172.233.228[.]93/patch", using the bash shell.
"The attackers have meticulously managed an access control list for the command-and-control server to ensure only the targeted device can access it," explains Unit 42’s analysis.
The vulnerable versions are PAN-OS 10.2, 11.0, and 11.1, and fixes for these versions are expected by April 14, 2024. The vendor will implement hotfixes by Sunday with the release of the following versions:
Products like Cloud NGFW, Panorama appliances, and Prisma Access are not affected. An overview of the impact can be seen in the table below:
Meanwhile, a threat researcher named Yutaka Sejiyama recently noted on X, "Scans show there are currently 82,000 exposed devices online that might be vulnerable to CVE-2024-34000, with 40% residing in the United States."
In response to the discovery, Volexity, which detected active exploitation of the flaw, identified a Python-based backdoor being delivered through these attacks.
"The script will then create another thread that runs a function called restore," said a spokesperson from Unit 42.
"The restore function takes the original content of the bootstrap.min.css file, as well as the original access and modified times, sleeps for 15 seconds and writes the original contents back to the file and sets the access and modified times to their originals."
These manoeuvres aim to erase traces of the intrusion, necessitating rapid exfiltration of data within 15 seconds before restoration of the files.
As Volexity pointed out, the threat actor has been able to create a reverse shell, download tools, infiltrate internal networks, and exfiltrate data, although the full scope of the campaign remains unclear.
The U.S. Cybersecurity and Infrastructure Security Agency (CISA) and Australia's Australian Cyber Security Centre (ACSC) are both responding to a critical vulnerability in Palo Alto Networks' PAN-OS software.
Australian organizations with a Palo Alto Threat Prevention subscription can defend against this CVE by activating Threat ID 95187, a measure introduced in the latest Applications and Threats content update.
The ACSC is actively monitoring the situation and is prepared to offer support and guidance as needed. Impacted organizations or individuals in need of assistance are encouraged to reach out through the hotline at 1300 CYBER1 (1300 292 371).
Meanwhile, CISA has included this vulnerability in its Known Exploited Vulnerabilities catalogue, compelling federal agencies to implement required patches by April 19 to prevent potential threats. Palo Alto Networks is expected to release a remedial update for this vulnerability by April 14.
Anthropic’s Fable 5 briefly gave Australia a rare look at Mythos-class cyber AI in action. Then US export controls shut access down, raising a harder question: if the model is too dangerous to leave America, are allies left safer, or simply more exposed?
The US-Iran conflict has triggered an unprecedented surge in cyberattacks. Between February 28 and March 20, DDoS attacks in the Middle East increased eightfold, with StormWall recording 2,000 to 3,000 attacks per minute at peak intensity.
ASIO’s $12.5 billion espionage warning is more than a tally of stolen secrets. It reveals a national digital crisis. With 24 major spy operations disrupted and identity systems exposed, Australia’s critical infrastructure and social services face a growing risk of collapse from unseen cyber threats.
Asia-Pacific faces escalating cyber threats: Qantas breach exposes 6M records, ransomware surges 57%, Japan doubles cybersecurity workforce by 2030. Singapore implements anti-scam laws, Indonesia establishes AI task force. Aviation sector under coordinated attack by Scattered Spider group.
Where cybersecurity meets innovation, the CNC team delivers AI and tech breakthroughs for our digital future. We analyze incidents, data, and insights to keep you informed, secure, and ahead. Sign up for free!
