Caja is a tool for safely embedding third party HTML, CSS and JavaScript in your website. - googlearchive/caja

From security leaders to strategic insights: a look back through a year of Security magazine.

 Cyberattacks are hitting businesses more often, fueled by automation and AI that accelerate the exploitation of vulnerabilities. Yet despite increasingly sophisticated techniques, experts say human mistakes, weak passwords, and poor access controls remain the biggest causes of successful breaches. While threats continue to evolve, people are still the weakest link in cybersecurity.  A recent report from Mandiant highlights how cybercriminal groups now operate through specialized teams. One group focuses on gaining access through phishing emails, malicious ads, or fake software updates, while another takes over to move through networks, steal data, or deploy ransomware. Attackers are also moving much faster. The average handoff time between criminal groups fell from more than eight hours in 2022 to just 22 seconds in 2025.  Vulnerabilities are increasingly exploited within days of disclosure, leaving organizations little time to patch systems before attacks begin. Cyber threats generally fall into two categories: financially motivated criminals seeking ransom payments or stolen data, and espionage-focused actors aiming for long-term, hidden access. While most intrusions are detected within about two weeks, cyber-espionage campaigns often remain unnoticed for more than three months.  Software vulnerabilities remain the leading attack vector, with technology and financial firms among the most targeted sectors. Researchers also observed a rise in voice-based social engineering, where attackers impersonate employees and contact IT help desks to bypass multi-factor authentication protections. Artificial intelligence is increasingly being used by threat actors for reconnaissance, phishing, and malware development. Some malicious tools even search compromised systems for AI-related credentials and resources.  However, researchers stress that AI is rarely the direct cause of breaches. Most incidents still stem from human error, weak security practices, misconfigurations, and excessive permissions. Ransomware attacks are evolving as well. Instead of only encrypting files, attackers now target backup systems, virtualization platforms, and recovery tools. By disabling recovery options, they increase pressure on victims to pay ransom demands. There are positive signs for defenders.  More organizations are detecting attacks internally through improved visibility, monitoring, and threat detection capabilities. Earlier discovery allows security teams to respond faster and reduce potential damage. Experts recommend stronger identity protection, continuous access verification, isolated backup environments, centralized login management, and behavior-based monitoring systems.  As cyber threats continue to accelerate, many security professionals believe identity security has become the new perimeter, making proactive defense more important than ever.

 Unexpectedly, Amazon Web Services’ Bedrock - built for crafting AI-driven apps - is drawing sharper attention from cybersecurity experts. Several exploit routes have emerged, threatening to reveal corporate infrastructure. Although the system smooths links between artificial intelligence models and company software, such fluid access now raises alarms. Because convenience widens exposure, what helps operations may also invite intrusion.   Eight ways into Bedrock setups emerge from XM Cyber’s analysis. Not the models but their access settings, setup choices, and linked tools draw attacker focus. Threats now bend toward structure gaps instead of core algorithms. How risks grow changes shape - seen here in surrounding layers, not beneath.  What makes the risk stand out isn’t just technology - it’s how Bedrock links directly to systems like Salesforce, AWS Lambda, and Microsoft SharePoint. Because of these pathways, AI agents pull in confidential information while performing actions across business environments. Operation begins once integration takes hold, placing automated units at the heart of company workflows.  A significant type of threat centers on altering logs. When attackers gain entry to storage platforms such as Amazon S3, they may collect confidential prompts - alternatively, reroute records to outside destinations, allowing unseen data transfers. Sometimes, erasing those logs follows, wiping evidence of wrongdoing entirely.  Starting differently each time helps clarity. Access points through knowledge bases create serious risks. Using retrieval-augmented generation, Bedrock pulls information from places like cloud storage, internal databases, or SaaS tools. When hackers obtain entry to those systems - or the login details tied to them - they skip past the AI completely. Getting in this way lets them grab unfiltered company data. Movement across linked environments also becomes possible.  Though designed to assist, AI agents may become entry points for compromise. When given broad access, bad actors might alter an agent's directives, link destructive modules, or slip corrupted scripts into backend systems. Such changes let them perform illicit operations - editing records or generating fake profiles - all while appearing like normal activity. What seems like automation could mask sabotage beneath routine tasks. One risk involves changing how workflows operate.  When Bedrock Flows get modified, information may flow through harmful components instead of secure paths. In much the same way, tampering with safeguards - those filters meant to block unsafe content - opens doors to deceptive inputs. Without strong barriers, systems face higher chances of being tricked or misused. Prompt management systems tend to become vulnerable spots. Because templates move between apps, harmful directions might slip through - reshaping how AIs act broadly, without needing new deployments, which hides activity longer.  Security teams worry most about small openings turning into big breaches. Though minimal, access might be enough for intruders to boost their permissions. One identity granted too much control could become a pathway inward. Instead of broad attacks, hackers exploit these narrow points deeply. They pull out sensitive information once inside. Control over AI systems may shift without warning. Cloud setups face risks just like local networks do.  Although researchers highlight visibility across AI tasks, tight access rules shape secure Bedrock setups. Because machine learning tools now live inside core business software, defenses increasingly target system architecture instead of algorithm accuracy.

Bluetooth headphones and earbuds are everywhere, and we were wondering what attackers could abuse them for. Sure, they can probably do th...

A new attack dubbed 'SmartAttack' uses smartwatches as a covert ultrasonic signal receiver to exfiltrate data from physically isolated (air-gapped) systems.