Kirsten Doyle

This year’s AI Appreciation Day shines a light on the rising power of artificial intelligence in every field. Cybersecurity experts come together to discuss what AI has achieved, and the hurdles it still faces.   Cybersecurity experts share their views with Information Security Buzz:   Traditional Access Controls Fall Short   Rom Carmel, Co-founder and CEO at Apono, adds that unlike static on-prem environments, cloud infrastructure is distributed and dynamic, requiring real-time capabilities to manage access securely and efficiently. “As organizations scale and adopt multi-cloud architectures, traditional access controls often fall short, lacking the agility and context awareness needed to keep pace.” Carmel…

Threat actors have a new trick: hiding malicious JavaScript inside what looks like an innocent image, according to the Ontinue research team. A string of phishing campaigns is using SVG (Scalable Vector Graphics) files to smuggle browser redirects past traditional security tools. The result? Stealthy attacks, minimal user interaction, and victims who never see it coming. Images That Bite SVGs aren’t just pictures. They’re text-based XML files, which means attackers can slip JavaScript into them without raising alarms. In these campaigns, the SVG files include hidden scripts disguised within script tags, using a format that conceals the actual code content.…

The Cybersecurity and Infrastructure Security Agency (CISA) has issued a grave warning about a critical vulnerability affects railroad communication systems across the US.   The flaw, designated as CVE-2025-1727, can potentially enable bad actors to control train brakes remotely (radio-proximity, not global internet).  This vulnerability focuses on the End-of-Train and Head-of-Train protocols, collectively known as FRED.   These systems link trains in movement. This vulnerability stems from insecure authentication within the protocol. Attackers can exploit this by using software-defined radio to spoof brake control packets.  If exploited, the consequences could be dire. Unauthorized commands might cause sudden stops or brake failures. Such…

Four young people have been arrested after cyber attacks on some of the UK’s best-known retailers.  The National Crime Agency picked up two 19-year-old men, a 17-year-old boy, and a 20-year-old woman. Early morning raids. Homes searched. Devices seized.  All remain in custody.  The arrests follow coordinated attacks on M&S, Co-op and Harrods in April. The suspects are being questioned on suspicion of blackmail, money laundering, computer misuse, and organised crime.  Paul Foster, Deputy Director of the NCA’s National Cyber Crime Unit, said:  “Since these attacks took place, specialist NCA cybercrime investigators have been working at pace and the investigation…

Four linked vulnerabilities in OpenSynergy’s Blue SDK allow attackers to take over a vehicle’s infotainment system with a single click. The flaws affect major automotive brands, including Mercedes-Benz, Volkswagen, and Skoda.  The PCA Security Assessment Team discovered the issues while analyzing compiled Blue SDK binaries. They didn’t have source code. They didn’t need it.  The vulnerabilities affect the Bluetooth protocol stack used by many embedded systems in the automotive supply chain. Together, the flaws create a path to remote code execution in the operating system of the affected unit.  Four Vulnerabilities, One Attack Chain  The bugs span two components of…

A new vulnerability in Google Gemini for Workspace shows how AI can be turned into a silent accomplice.  A security researcher has uncovered a way to smuggle malicious commands into an email, hidden from the user’s view but faithfully executed by Gemini.   When the recipient clicks “Summarize this email,” Gemini parses the invisible instruction and inserts a phishing warning that appears to come directly from Google.  There are no links. No attachments. Just invisible code buried in the email body.  This indirect prompt injection (logged by 0DIN as submission 0xE24D9E6B) relies on HTML and CSS trickery. A few lines of…

A security lapse in McHire, McDonald’s chatbot-powered recruitment platform, exposed personal data from more than 64 million job applicants. The breach combined two fundamental flaws: default administrator credentials and an insecure direct object reference (IDOR) vulnerability. McHire, used by 90% of McDonald’s franchisees, lets candidates chat with a bot named Olivia, built by Paradox.ai, to apply for jobs, share personal details, and take a brief personality test.  Two independent security researchers, Ian Carroll and Sam Curry, uncovered the flaws during a brief, informal review after spotting user complaints about the chatbot’s erratic behavior on Reddit.   “123456” and You’re In   While…

Iran doesn’t need to fire missiles to be dangerous. It already sees the UK as a threat. And it’s acting accordingly. That’s the clear warning from Parliament’s Intelligence and Security Committee in its July 2025 report. Drawing on classified material, the report exposes Iran’s deep-rooted hostility to the West, and how that ideology is driving a wide range of hostile activity against the UK. From targeted cyber attacks and online espionage to physical threats against dissidents and journalists living in Britain, the Iranian regime is playing a long game. One shaped by survival instinct, historic grievances, and a willingness to…

Europe’s manufacturing sector is under siege. A new report from KnowBe4, released this week, lays bare the growing cyber vulnerabilities facing the continent’s increasingly digitalised production lines. Titled Securing Manufacturing’s Digital Future, the report paints a stark picture: as factories adopt smarter, more connected systems, they are becoming prime targets for ransomware gangs and social engineering attacks. Digital progress has its price Manufacturers are pouring investment into automation and smart infrastructure. But the result is a sprawling attack surface, one that blends IT and operational technology, and opens the door to exploitation.  “Every minute of disruption costs money,” the report…

A high-severity vulnerability in the ServiceNow platform could have exposed vast amounts of sensitive data to low-privileged or even anonymous users. Researchers at Varonis Threat Labs discovered the issue, dubbed Count(er) Strike, which exploits a flaw in how the system displays record counts, offering attackers a quiet but powerful method of data inference and exfiltration. “Any user in an instance could exploit this vulnerability, even those with minimal privileges and no assigned roles,” said Varonis. “All they needed was access to a single misconfigured table.” The exposure risk was broad. ServiceNow is used by 85% of the Fortune 500 to…