Search
Close this search box.
Search
Close this search box.

Armilla: Providing Warranties on Third-party AI Models for Corporate Customers

Image Credits: iStock.com/Thinkhubstudio
By: Headliners News / February 29, 2024

Meet Armilla, a unique startup spearheaded by entrepreneurs Karthik Ramakrishnan and Dan Adamson, has entered the AI landscape with a distinctive proposition — offering warranties on third-party AI models to address issues of trust and quantify risk. With Ramakrishnan’s background as a co-founder of two “AI-first” companies, Gallop Labs and Blu Trumpet, he recognized that a lack of trust was hindering widespread AI adoption in enterprises.

The unique aspect of Armilla’s approach lies in its benchmark assessments. Given that many AI models remain opaque or guarded behind licenses, subscriptions, and APIs, Armilla takes a meticulous approach to assess and quantify the risk associated with these models. The company’s in-house assessment technology tests for hallucinations, bias, fairness, general robustness, and security, drawing insights from a range of theoretical applications and use cases.

The underlying premise is to provide confidence to enterprises procuring third-party AI solutions by thoroughly evaluating the models they intend to deploy. Armilla’s comprehensive assessments cover various facets, including Key Performance Indicators (KPIs), processes, performance, data quality, and both qualitative and quantitative criteria. The assessments align with global AI regulatory frameworks, considering factors such as hallucinations, racial and gender bias, fairness, and overall robustness.

The ultimate differentiator for Armilla is its warranty offering, which acts as a safety net for enterprises investing in AI models. The warranty kicks in when a model passes the rigorous assessment, ensuring that Armilla will reimburse the buyer for any fees associated with using the model. This not only adds a layer of confidence but also aligns with Armilla’s commitment to ethical AI practices.

While warranties and policies for AI have been explored previously, Armilla’s approach is positioned as unique within the AI landscape. The company debuted its coverage in late 2023, supported by leading carriers such as Swiss Re, Greenlight Re, and Chaucer.

As the AI regulatory landscape evolves globally, Armilla claims to keep pace by regularly updating its assessments to align with emerging regulations. The company emphasizes that it refuses to conduct assessments for AI models that fall into prohibited categories or are banned by regulatory bodies.

Armilla’s client base has reportedly been growing at an impressive rate, doubling month over month since Q4 2023. The company’s clientele spans enterprises and third-party AI vendors, with both segments leveraging Armilla’s services for different purposes.

With $4.5 million recently secured in a seed round led by Mistral, along with participation from prominent names such as Greycroft, Differential Venture Capital, and Mozilla Ventures, Armilla plans to utilize the funds to expand its existing warranty offerings and introduce new products. The funding round takes Armilla’s total raised capital to $7 million, reflecting the growing confidence in this startup‘s unique approach to mitigating risks associated with third-party AI models.

Ramakrishnan remains optimistic about Armilla’s role in shaping the future of AI risk mitigation, noting that insurance will play a big role in addressing AI risk, and Armilla is at the forefront of developing insurance products that will allow companies to safely deploy AI solutions. The company’s commitment to creating a comprehensive approach to AI risk assessment and providing a safety net through warranties positions it as a significant player in the evolving landscape of AI governance and ethics.

What do you think?

3 People voted this article. 3 Upvotes - 0 Downvotes.
Please Share This