ESET announced a €40 million investment to expand its B2B cybersecurity offerings with capabilities designed to defend AI workflows and agents. The company says that while AI is welcomed in workplaces as an efficient virtual colleague, it has simultaneously become a fast and ruthless ally for cybercriminals.
Why the risk is growing
The rise of AI has brought practices such as "vibe coding," enabling people without traditional programming skills to produce websites, apps or browser extensions simply by specifying requirements to an AI. However, studies cited by ESET indicate that between 40% and 70% of AI‑generated code contains some form of security flaw or hidden vulnerability, depending on language, task and model used.
Employees and enthusiastic ‘‘vibe coders’’ frequently connect AI agents to sensitive data or feed confidential company information into AI chatbots to increase efficiency, often without understanding the privacy and security risks. ESET warns that such incidents can have severe financial, legal and reputational consequences for organizations.
Phishing has become both faster and more convincing: crafting a persuasive, personalized phishing email previously took about 16 hours; with AI this can be accomplished in roughly five minutes.
AI‑driven malware and ransomware
ESET Research identified the first AI‑based ransomware, PromptLock, and an Android malware that exploits generative AI, PromptSpy. In its latest Threat Report, ESET inspected 900,000 public AI capabilities (skills); of these, 25,000 were deemed suspicious and more than 3,000 were explicitly malicious. These agents were designed for data exfiltration, downloading malware, and covertly altering agent behavior.
ESET underscores that AI is no longer only a tool for attackers: it can autonomously discover zero‑day vulnerabilities and generate exploit code.
New ESET AI defense modules and product updates
To help companies address these emerging threats, ESET expanded the ESET PROTECT platform with new B2B features focused on AI workflows and agents. The expanded portfolio includes two key modules:
- ESET AI agent security: monitors what AI agents can access and download, preventing malicious components from entering corporate systems.
- ESET AI behavior analysis: observes how AI tools and agents behave on endpoints during routine operation, detecting suspicious or unexpected actions early, before they escalate into incidents.
ESET PROTECT also integrates ESET LiveCortex, a correlation and incident response engine. The company’s generative AI cybersecurity assistant, formerly called AI Advisor, has been renamed ESET LiveAI to better reflect its place in the product portfolio and its capability to provide real‑time context for remediation.
Michal Jankech, Vice President for Enterprise, SMB and Managed Service Provider segments at ESET, commented: “As cybersecurity enters the era of autonomous artificial intelligence, ESET is supporting organizations with a renewed B2B portfolio to address new AI‑based threats. This new lineup — together with our €40 million AI investment — reflects our commitment to closing AI security gaps and ensuring defensive considerations remain central to innovation.”
ESET said the new protections will appear first in the business portfolio and will be rolled out to consumer products soon after.
Hungarian survey on workplace AI use
In Hungary, Sicontact Kft., a distributor of ESET products, conducted a survey examining how much Hungarian employees trust AI, whether managers are aware that staff may process sensitive company data with AI, and how AI changes users’ security decisions. The industry report discusses the IT security, employer and mental‑health responsibilities that AI use imposes on organizations. The white paper “MI a baj? Mesterséges intelligencia, IT‑biztonság és mentális egészség” is available for download at www.miabaj.hu.
