ESET Previews New AI Security Features to Secure Chatbot Communications and AI Workflows

• Showcased at RSAC 2026, ESET’s upcoming AI security measures will defend the total AI dialog movement by scanning each prompts and responses to scale back information publicity and compliance dangers.
• Constructed as a browser safety function, it is going to defend in opposition to malicious hyperlinks, scripts and content material generated by LLMs and forestall add of delicate and confidential information into public AI methods. 
• ESET additionally introduced its new endpoint safety capabilities designed to safe private AI assistants from AI provide chain assaults whereas making a free, public ESET AI Skills Checker to detect dangerous and malicious conduct of AI expertise earlier than deployment.

SAN DIEGO, April 02, 2026 (GLOBE NEWSWIRE) — ESET, a worldwide chief in cybersecurity, at present introduced upcoming AI safety capabilities designed to safe how workers work together with AI instruments. Demonstrated at RSAC 2026 and set to launch later this 12 months, the brand new function will broaden visibility within the ESET PROTECT Platform to research rising dangers tied to on a regular basis AI utilization and agentic AI adoption throughout the enterprise.

“As corporations rely extra on AI for productiveness and automation, they face rising dangers round delicate information publicity, compliance violations, and deceptive outputs,” stated Juraj Jánošík, ESET Director of Synthetic Intelligence. “Agentic AI is shifting the safety battlefield again to the endpoint. ESET has spent over 30 years constructing main endpoint safety powered by AI and machine studying, so we’re uniquely positioned to assist organizations safe this subsequent wave of AI proper the place it begins.”

As AI instruments develop into embedded in on a regular basis workflows, many workers are utilizing open cloud chatbots with out IT oversight, creating “shadow AI” dangers and exposing delicate information reminiscent of inside paperwork, API keys, secrets and techniques, and credentials. ESET addresses this by means of numerous applied sciences that get as near the supply as attainable, considered one of which is a safe browser expertise that intercepts AI interactions and analyzes each prompts and responses in actual time—serving to forestall information publicity and detect malicious or deceptive content material earlier than it impacts customers.

In demonstrations at RSAC 2026, the brand new AI safety function flagged malicious URLs submitted by means of chatbot prompts, logging exercise on the endpoint and surfacing it within the ESET PROTECT Platform for investigation. The identical strategy applies to immediate injection makes an attempt, scripts, and delicate information inputs, enabling organizations to dam or monitor exercise based mostly on coverage. Safety groups will achieve visibility into how AI instruments are used throughout the group by means of ESET PROTECT Platform logging, serving to them examine dangers and implement insurance policies extra successfully.

As organizations broaden their use of agentic AI instruments, the assault floor extends past chatbot interactions to incorporate rising “AI provide chain” dangers. These embrace compromised AI frameworks and instruments, reminiscent of trojanized parts in extensively used libraries like LiteLLM, in addition to autonomous brokers like OpenClaw that may execute actions on a system with restricted oversight. ESET has already been defending its clients from supply-chain assaults by means of compromised libraries delivered through commonplace repositories however is noting an increase in most of these assaults and stays dedicated to additional analysis and growth referring to AI instruments.

As a part of its broader AI safety innovation, ESET launched a free ESET AI Expertise Checker at RSAC 2026. Obtainable to non-ESET clients and constructed on the identical expertise as ESET’s endpoint safety merchandise and ESET LiveGuard, the scanner analyzes AI expertise for hidden directions, malicious code, and dangerous conduct utilizing multi-layered inspection and cloud-based sandboxing. It is usually presently obtainable as a built-in function for present ESET Endpoint customers.

For greater than 30 years, ESET has pioneered light-weight, excessive‑efficiency endpoint safety powered by machine studying and synthetic intelligence. These new capabilities prolong that basis by serving to organizations defend in opposition to at present’s quickly shifting menace panorama, the place cybercriminals more and more harness AI to scale assaults, goal workers and automate subtle social engineering.

As the one devoted cybersecurity member of the Agentic AI Basis (AAIF), ESET can also be working to safe rising AI agent communication protocols by means of collaboration with trade leaders like OpenAI, Amazon, Microsoft, and Anthropic. Collectively, the group is working to determine trusted requirements, safe protocol designs, and greatest practices for AI agent interoperability.

Be taught extra about why companies select ESET at https://www.eset.com/us/business/why-eset/

About ESET

ESET® offers cutting-edge cybersecurity to stop assaults earlier than they occur. By combining the ability of AI and human experience, ESET stays forward of rising world cyberthreats, each identified and unknown—securing companies, vital infrastructure, and people. Whether or not it’s endpoint, cloud, or cellular safety, our AI-native, cloud-first options and companies stay extremely efficient and straightforward to make use of. ESET expertise consists of strong detection and response, ultra-secure encryption, and multifactor authentication. With 24/7 real-time protection and robust native help, we hold customers secure and companies operating with out interruption. The ever-evolving digital panorama calls for a progressive strategy to safety: ESET is dedicated to world-class analysis and highly effective menace intelligence, backed by R&D facilities and a robust world companion community. For extra data, go to http://www.eset.com or observe our social media, podcasts, and blogs.