Advanced Computing in the Age of AI | Monday, June 24, 2024

Galileo Introduces Protect, a Real-Time Hallucination Firewall to Safeguard Enterprise Generative AI 

SAN FRANCISCO, May 2, 2024 – Galileo, a leader in developing generative AI for the enterprise, today announced the release of Galileo Protect, a real-time GenAI firewall solution that proactively intercepts malicious inputs and stops harmful responses, such as hallucinations, in Large Language Model (LLM) applications.

Building on the success of Galileo Evaluate and Galileo Observe, the company’s evaluation and real-time monitoring solutions, Protect uses Galileo’s suite of research-backed metrics to protect LLM applications in real time, all from one easy-to-use, centralized user interface for AI and compliance teams.

“The rapid adoption of AI has introduced a new set of safety and compliance risks that need to be managed by enterprises. Organizations are struggling to safeguard AI without impacting user experience; nobody wants to talk to a chatbot that takes two seconds to respond,” said Galileo CTO and Co-Founder Atin Sanyal. “Galileo has created guardrails that can identify and block malicious activities in real-time, allowing enterprises to enhance AI safety without impacting performance or user experience. And without incurring significant cost. This is built on the back of innovative breakthroughs by Galileo Research.”

Before Galileo Protect, managing LLM compliance and security was challenging and tedious, with no way to easily secure AI applications from malicious inputs or protect end users from hallucinatory or unsafe responses. With Galileo Protect’s suite of high-accuracy, low-latency metrics, AI practitioners have the ability to bring predictability and control to their GenAI applications using real-time request and response interception.

The solution removes the most persistent problems faced by AI engineers when setting up firewalls for LLM applications by allowing them to systematically analyze and iterate firewall configurations during testing and evaluation.

Key Galileo Protect features include:

  • Proactive protection from bad inputs: Hardens GenAI system against malicious activities, such as prompt attacks, offensive inputs, and more.
  • Block harmful model response: Controls application outputs to protect users from hallucinations, data leakage, and off-brand responses.
  • High-accuracy metrics: Production-ready research-backed metrics that precisely detect issues, anomalies, and security threats.
  • Millisecond latencies and cost efficiency: Rapid interception without slowing the end-user experience or frequent and costly API calls.
  • Rule configuration and management from a centralized UI: Allows collaboration across non-technical and technical teams to configure, monitor, manage, and edit Protect Rules.

Galileo Protect has helped multiple enterprises accelerate AI development, reduce response times, and improve AI security, safety, and governance with Galileo’s powerful suite of guardrail metrics, baked-in controls, and customization.

“GenAI has been a major strategic initiative for us, and we’ve leveraged it extensively in our ongoing pursuit for efficiencies and enhanced end-user experience,” said Darrel Cherry, Distinguished Engineer at Clearwater Analytics. “Using Galileo Observe, we’ve significantly enhanced our visibility and reduced response times to any potential issues with LLMs from days to minutes. Galileo Protect takes this a step further by allowing us to automatically monitor and intercept AI responses in real-time, enabling us to provide guardrails around our AI products and bring them to customers faster."

Read the latest Galileo blog to learn more about the many ways Galileo Protect and the suite of Galileo products can help your AI team productionize GenAI in your enterprise, accomplishing fast, safe, and compliant results.

About Galileo

San Francisco-based Galileo is the leading platform for enterprise GenAI evaluation. The Galileo platform, powered by research-backed metrics, supports AI teams across the development lifecycle, from building and iterating to monitoring and protection. Galileo is used by AI teams from startups to Fortune 100 companies.

Source: Galileo