- Qualys (QLYS, Financial) enhances TotalAI solution, focusing on AI security from development to deployment.
- New features include on-premises LLM scanner and detection of 40 attack scenarios.
- 72% of CISOs express concerns over AI-related security risks, addressed by TotalAI updates.
Qualys, Inc. (QLYS), a leader in cloud-based IT security, has announced significant updates to its TotalAI solution. The enhancements aim to secure organizations' Machine Learning Operations (MLOps) pipelines from development to deployment, addressing the growing security concerns surrounding AI technologies.
The TotalAI update introduces an internal Large Language Model (LLM) scanner for on-premises testing, enhancing security during development without external exposure. It now detects 40 different attack scenarios, including advanced jailbreak techniques, prompt injections, multilingual exploits, and bias amplification.
72% of Chief Information Security Officers (CISOs) have expressed concerns about potential security breaches arising from generative AI solutions. Qualys TotalAI helps bridge the gap between innovation and secure implementation, ensuring only vetted models are deployed into production.
"As AI becomes core to business innovation, security can no longer be an afterthought," remarked Tyler Shields, principal analyst at Enterprise Strategy Group. TotalAI supports a "shift-left" approach, integrating security into Continuous Integration and Continuous Deployment (CI/CD) workflows to protect sensitive models behind corporate firewalls.
Additionally, TotalAI's enhanced multimodal threat detection defends against attacks targeting LLM outputs via images, audio, and video, empowering organizations to stay ahead of AI threats. Sumedh Thakar, president and CEO of Qualys, emphasized TotalAI's capacity to protect AI workloads across all stages—enabling businesses to innovate securely.
Qualys TotalAI is available now, with a 30-day trial option to explore its comprehensive security capabilities. For more information, visit Qualys' website.
