ESET adds ChatGPT link and script checks, with Gemini and Copilot support still to come

ESET has announced an AI Security layer in updates to ESET HOME Security and ESET Small Business Security. The October 27 release describes checks around links and scripts returned by AI tools, plus monitoring of agent downloads and behavior.

The assistant‑specific feature, AI Conversation Security, initially names ChatGPT. ESET says it scans URLs in responses and flags links associated with phishing, malicious or potentially unwanted sources before a user follows them. It also describes scanning scripts generated by an LLM. These are the company’s descriptions of the checks; they do not establish that every suspicious link or script will be detected, or that a response without a warning is correct.

ESET says Gemini and Copilot integrations are coming soon. The announcement does not provide their release dates. A ChatGPT integration should therefore not be read as current coverage for every assistant or every surface through which someone might use an AI model.

The release describes AI Agent Security as monitoring what agents access and download before execution, including files, scripts and external URLs. A separate AI Behavioral Monitoring component is described as observing agents’ actions for unusual activity, suspicious access attempts, unexpected behavior patterns and download sources.

Those descriptions concern different points in a workflow: an item an agent is about to use, and the actions an agent takes over time. They are not a published guarantee that an agent cannot perform an unwanted action. OpenTools has not installed the update, run a malicious‑sample test or verified detection and prevention behavior.

ESET also describes an AI Skills Checker as a free online tool for scanning add‑ons for signs of malicious activity and risky behavior before installation. That free checker is a separate offering from the named consumer and small‑business subscriptions. Its description does not make the paid products free, establish every platform entitlement, or turn a scan result into permission to install or run untrusted code. No skill was submitted or executed for this report.

The announcement also discusses ESET Mobile Security for iOS. Its body uses current‑availability language, but the footnote says the app is pending Apple App Store approval, will be available soon and may vary by region. That qualification prevents treating the body wording as proof of a completed universal iOS release.

The iOS section lists its own link, device‑setting and connection features. It does not establish that every newly announced AI Conversation, Agent or Behavioral Monitoring function runs on iOS. The retained release also does not provide a complete feature‑by‑plan, operating‑system, browser or version matrix for the AI layer.

The concrete product update is therefore a set of ESET‑described checks, with ChatGPT named first and other assistant integrations still forthcoming. Availability for a particular device and subscription needs to be established from that product’s actual entitlements rather than inferred from the overall announcement.

Original source‑based figure by OpenTools Team, based on ESET’s announcement. ChatGPT is named first for conversation checks; Gemini and Copilot are forthcoming. The separate iOS app remains subject to approval and regional availability. Feature and device entitlements are not interchangeable. This is an explanatory diagram, not a product screenshot or detection test.