AI Research

Driving real-world impact

See Demo

Publications

pub1

No free lunch with Gaurdrails

Benchmarks show stronger guardrails improve safety but can reduce usability. Paper proposes a framework to balance the trade-offs — ensuring practical, secure LLM deployment.

Read Paper
pub2

Investigating Implicit Bias in LLMs

A study of 50+ models reveals that bias persists — and sometimes worsens — in newer models. The work calls for standardized benchmarks to prevent discrimination in real-world AI use.

Read Paper
pub3

VERA: Validation & Enhancement for RAG

VERA improves Retrieval-Augmented Generation by refining retrieved context and output, reducing hallucinations and enhancing response quality across open-source and commercial models.

pub4

Fine-Tuning, Quantization & Safety

Fine-tuning increases jailbreak vulnerability, while quantization has varied effects. Our analysis emphasizes the role of strong guardrails in deployment.

pub5

SAGE-RT Synthetic Red Teaming

SAGE enables scalable, synthetic red-teaming across 1,500+ harmfulness categories — achieving 100% jailbreak success on GPT-4o and GPT-3.5 in key scenarios.

Such research advancements fuel our security platform and power the

LLM Safety & Security Leaderboard

The most comprehensive benchmark for model safety in the industry.
LLM Safety Leaderboard

AI Guardrail Benchmark Studies

Our research goes beyond just publications – it’s been applied to real-world benchmark studies to evaluate the security and performance of leading AI guardrails. These comparative tests provide practical insights into how guardrails perform under real attack scenarios, driving measurable improvements in AI safety.

Enkrypt AI, Guardrails AI, and Protect AI LLM Guard

Read Blog

Enkrypt AI vs Azure Content Safety vs Amazon Bedrock Guardrails

Read Blog

Enkrypt AI, IBM Granite, Azure AI, Prompt Shield, and Amazon Bedrock Guardrails

Read Blog
Graphic Element-Horizontal Lines

Building Safer AI from the Ground Up:
Securing LLM Providers

Enkrypt AI tests over 100 leading foundation models - including from AI21, DeepSeek, Databricks, and Mistral - to strengthen the safety of their LLMs without compromising performance.
Insecure Code
CBRN
Health
Toxicity
Race
Biases
Harmful Content
Religion
Gender
Hate Speech
Discrimination
Suicide & Self-harm
Harmful Content
Sexual Content
Gender
Insecure Code
Health
Toxicity
50,000+ tests
CBRN
Regulated or Controlled Substances
Guns & Illegal Weapons
Criminal Planning
Harmful Content
Biases
We conduct more than 50,000 dynamic red-teaming evaluations per model, spanning critical risk categories: bias, insecure code, CBRN threats, harmful content, and toxicity. This rigorous testing ensures our insights are among the most trusted in the industry.