Lasso Security announced
the launch of RapidClassifier, its patent-pending
technology designed to accelerate the security of generative AI deployments and
mitigate latency issues.
Adding in-line security to Generative AI adds significant
complexity and latency which puts organizations
in a dilemma: either slow down their applications to accommodate protections or
leave themselves exposed.
Organizations are rapidly adopting Generative AI - according to Deloitte, two thirds of companies they recently surveyed are
increasing investments in the technology as they have seen strong value to
date. Generative AI is expected to be immediate and natural. Responding to user
inputs and processing data at the seconds level is the standard and users
expect their experience to match this.
Lasso's RapidClassifier is the first framework designed to close this gap. By
enabling Lasso's custom security policies based on LLMs to run in under 50 milliseconds, it ensures that the speed of protection matches the
speed of Generative AI itself.
"Many companies are hesitant to add security and
guardrails to Generative AI applications due to latency concerns. With
RapidClassifier, we're able to mitigate these concerns and empower companies to
unlock genAI's potential, bridging the gap between speed and security," said
Ophir Dror, CPO & Co Founder, Lasso Security. "Generative AI is redefining
what technology can do, and security must evolve to keep pace. RapidClassifier
is more than a speed boost-it's a paradigm shift in how flexible and simple
security can be. By combining dynamic adaptability with lightning-fast
execution, it offers a new way to think about security in the age of generative
AI."
Key Benefits of RapidClassifier:
- Custom Policies,
Optimized for Speed: RapidClassifier
allows organizations to implement highly specific, flexible policies-and
ensures they execute in under 50 milliseconds, eliminating delays.
- Proactive Protection: By running policies in real-time, RapidClassifier
stops potential threats before they can escalate, bridging the gap between
detection and prevention.
The integration of large language models (LLMs) in business
operations promises enhanced productivity and innovation. However the dynamic, undeterministic nature of large
language models (LLMs) means traditional security solutions, reliant on
pre-defined patterns and rigid frameworks, are ill-equipped to handle the
constantly shifting outputs and interactions of generative AI. Organizations must address these risks to safeguard data and
maintain compliance.
Working at the forefront of Generative AI security, Lasso
protects businesses that are integrating LLMs into their operations,addressing
challenges in three critical areas : content anomaly detection, privacy and
data protection, and AI application security. Lasso actively monitors large
language model outputs, ensuring compliance with organizational standards and
preventing data leakage, while providing robust defense to guarantee the
security and integrity of large language model-based applications.