Protect AI announced it has acquired Laiyer AI. With the
acquisition, Protect AI will be offering a commercial version of Laiyer
AI’s open source LLM Guard with expanded features, capabilities, and
integrations within the Protect AI platform. LLM Guard is freely
available today, and an industry leading open-source project for
protecting large language models (LLMs) against security threats, misuse
and prompt injection attacks, while also providing tools to manage risk
and compliance needs.

OpenAI’s GPT-4 and other LLMs are revolutionizing AI, and excel at
understanding and generating human language. Their adoption spans
various sectors, including customer service, healthcare, and content
creation, driving the market’s growth from USD 11.3 billion in 2023 to
an expected USD 51.8 billion by 2028, according to multiple industry
analysts. This growth, fueled by the demand for applications like
chatbots and virtual assistants, positions LLMs as key tools for
businesses seeking to leverage textual data for competitive advantage.
However, security and misuse concerns are limiting wider adoption among
major companies.

“Protect AI is thrilled to announce the acquisition of Laiyer AI’s team
and product suite, which significantly enhances our leading AI and ML
security platform. These new capabilities will empower our customers in
automotive, energy, manufacturing, life sciences, financial services,
and government sectors to develop safe, secure GenAI applications,” said
Ian Swanson, CEO of Protect AI. “Our industry-leading platform now
boasts advanced features and filters for governing LLM prompts and
responses, elevating the end-user experience and reaffirming our
commitment to safeguarding Generative AI applications,”

In 2023, the OWASP Top 10 for LLM Applications
spotlighted the unique security risks associated with deploying Large
Language Models that business leaders should understand. Key risks
include prompt injections, training data poisoning, and supply chain
vulnerabilities. A notable concern is Prompt Injection Vulnerabilities,
where attackers can manipulate LLMs through crafted inputs, leading to
data exposure or decision manipulation. These attacks can be direct, via
the LLM’s input, or indirect, through tainted data sources, and often
bypass detection due to the implicit trust in LLM outputs. With upcoming
regulations on LLMs, it’s vital to safeguard against such malicious
activities and harmful responses to maintain corporate integrity and
security.

Laiyer AI LLM Guard is a groundbreaking security solution for addressing
the challenges associated with deploying LLMs. Unlike many
closed-source, untested options prevalent in the market, LLM Guard
offers a transparent, open-source alternative that bolsters confidence
in deploying LLMs at an enterprise scale. This innovative tool is
designed to enhance the security of LLM interactions, supporting both
proprietary and third-party models.

LLM Guard’s core features include the detection, redaction, and
sanitization of inputs and outputs from LLMs, effectively mitigating
risks such as prompt injections to personal data leaks. These features
are integral to preserving LLM functionality while safeguarding against
malicious attacks and misuse. Moreover, LLM Guard integrates seamlessly
with existing security workflows, offering observability tools like
logging and metrics. This positions Laiyer AI at the forefront of
providing essential security solutions, enabling developers and security
teams to deploy LLM applications securely and effectively.

“There’s a clear need in the market for a solution that can secure LLM
use-cases from start to finish, including when they scale into
production. By joining forces with Protect AI, we are extending Protect
AI’s products with LLM security capabilities to deliver the industry’s
most comprehensive end-to-end AI Security platform,” said Neal Swaelens
and Oleksandr Yaremchuk, Co-founders of Laiyer AI.

LLM Guard exemplifies price/performance leadership in the enterprise
security sector for LLMs. The innovative solution balances latency,
cost, and accuracy, boasting an impressive scale of adoption with over
13,000 library downloads and 2.5 million downloads of its proprietary
models on HuggingFace in just 30 days. LLM Guard’s performance is
enhanced by a 3x reduction in CPU inference latency, enabling the use of
cost-effective CPU instances instead of expensive GPUs without
compromising accuracy. LLM guard is a leader in the field, reinforced by
its status as the default security scanner for Langchain and several
other leading global enterprises.