Check Point updates Quantum firewall to secure AI use

Check Point Software has released a new version of its Quantum firewall software that targets security risks arising from the rapid adoption of artificial intelligence in enterprises.

The R82.10 release introduces 20 new features across network, identity and AI security. It focuses on hybrid environments that span data centres, branch offices and multiple public clouds.

Check Point said the update responds to growing pressure on security teams as businesses roll out AI tools and large language model projects across their operations.

"As organisations embrace AI, security teams are under growing pressure to protect more data, more applications and more distributed environments," said Nataly Kremer, Chief Product Officer, Check Point Software Technologies. "R82.10 helps enterprises shift to a prevention-first model by unifying management, strengthening Zero Trust and adding protections that support safe, responsible AI adoption and development."

Many enterprises now connect AI services to existing systems and data stores. This raises concerns over data leakage, model abuse and new types of automated attacks.

Check Point said the new software focuses on four operational areas. These areas are AI adoption, hybrid mesh network security, threat prevention and platform unification.

AI oversight

The company has added functions that track and control the use of generative AI tools. The software can detect unauthorised AI services in network traffic. It adds specific visibility for tools such as ChatGPT, Claude and Gemini.

The release also includes monitoring for use of the model context protocol. This protocol supports AI workflows that combine different tools and data sources. The monitoring feature aims to reduce the risk of unmanaged AI agents accessing sensitive systems.

Check Point said these controls give security teams more detail on which AI services employees use and how data flows through those services.

Hybrid mesh focus

R82.10 targets organisations that manage many sites and clouds. The software links on-premise firewalls and secure access service edge, or SASE, services through centralised internet access policies.

This approach applies the same rules to traffic from branches, remote workers and cloud workloads. It reduces the need for separate policy sets in each environment.

The release also adds features for identity and device posture checks. These checks support Zero Trust models at larger scale. The software uses stronger validation of users and devices before it allows access to applications.

Threat prevention

Check Point has added phishing protection that does not require HTTPS inspection. This method can detect malicious sites and links without decrypting traffic content.

The company said this feature suits environments that must keep encrypted traffic private for regulatory or privacy reasons.

R82.10 also includes what Check Point calls adaptive intrusion prevention. The system adjusts alerts based on threat context. It aims to reduce alarm fatigue in security operations centres.

New Threat Prevention Insights highlight configuration weaknesses and posture gaps. The system flags issues before attackers can exploit them. Security teams can then prioritise remediation work based on these findings.

Unified platform

The update deepens Check Point's integration strategy. The company said R82.10 supports more than 250 third-party integrations.

Customers can feed endpoint posture data from existing providers into Quantum firewall policies. This data can include device health, operating system status and security agent information.

The firewall then uses these signals in identity-based rules. This approach tightens access controls and supports finer-grained Zero Trust enforcement.

Industry analysts note that AI change is reshaping network security strategies. "With the efficiency gains promised by AI, security professionals cannot slow down business innovation or risk being excluded," said Frank Dickson, Group Vice President, Security & Trust, IDC. "The benefits of innovation do not negate the looming security threat being introduced by AI. Enterprises need to reduce risk, unify controls and stay ahead of sophisticated malicious actors. Check Point's approach of embedded AI security into the network stack is an appropriate approach to quickly improve an organisation's AI security posture."

AI stack link

R82.10 forms part of Check Point's broader AI security portfolio. The firewall software integrates with the company's Infinity Platform and its open integration framework.

Check Point said the platform approach lets organisations apply the same policies across on-premise networks, cloud workloads and remote users.

The new firewall release also connects with AI security technology from Lakera. Check Point acquired Lakera earlier this year. Lakera's tools focus on securing AI models during training and inference.

Check Point said this combination offers protection for AI workloads from data ingestion through to model output inspection.

Channel partners see demand for such integrated approaches. "Check Point continues to deliver AI Security innovations at the exact moment customers need them," said Chris Konrad, Vice President, Global Cyber, World Wide Technology. "Their AI-driven security capabilities help organisations safeguard their businesses from the latest cyber threats, while providing enterprise-grade protection for sensitive AI workloads from model training to inference without compromising performance."

Check Point plans further development of its AI security features across the Quantum and Infinity product lines as customers expand AI projects across hybrid networks.