The development of AI is arguably one of the most significant human achievements since splitting the atom. However, as many who have used a Large Language Model [LLM] will attest, the responses can, at times, be unexpected. Useful and fun though they are, they cannot be reliably trusted when asked to perform important tasks where accuracy is critical. Current LLMs are prone to hallucinate and produce unexpected outcomes which could prove catastrophic in certain circumstances.
Y-Combinator-backed atla, an AI safety startup specialising in developing safety protocols for generative AI models in critical sectors like legal, financial, and medical applications, has successfully raised $5M in a Seed Round led by Creandum, a European early-stage investor and early backer of Spotify, Klarna, iZettle, Kahoot, Vivino, depop, and Trade Republic. This funding round also included participation from existing investors, Y Combinator and Rebel Fund whose partners include Reddit’s CEO Steve Huffman, Cruise co-founder Daniel Kan, Instacart co-founder Max Mullen, Scribd co-founder Trip Adler, and Sebastian Mejia, co-founder of Rappi.
Pioneering AI safety measures
London and San Francisco-based, atla, fresh out of Y Combinator’s latest accelerator batch, state their core mission is to establish secure AI systems tailored for high-stakes applications. This is to mitigate the inherent risks of errors and unexpected outcomes associated with large language models. The company’s innovative guardrails aim to enhance truthfulness, diminish harmfulness, and bolster reliability for generative AI models.
atla’s initial technology manifestation is an AI assistant developed in collaboration with legal teams at Volkswagen and N26. This AI assistant leverages atla’s research capabilities to furnish legal teams with synthesised answers derived directly from highly trusted legal sources in response to legal inquiries.
Leadership and expansion plans
Founded in January 2023, atla’s leadership comprises co-founders, Maurice Burger, an AI startup veteran who dropped out of his MBA at Harvard to pursue a degree in Computer Science, and Roman Engeler, who brings extensive experience from his research endeavours at ETH Zurich and Stanford, followed by leadership roles in two startups backed by Switzerland’s largest corporate venture builder.
The injection of $5M in funding is geared towards scaling atla’s product development and team expansion. Currently, atla is actively recruiting for technical positions to fortify its London-based team as part of its growth strategy.
Vision for safer AI systems
Maurice Burger expressed the company’s commitment to advancing the reliability of large language models, especially in high-stakes scenarios. The aspiration is to create AI systems that function consistently and reliably in critical situations, ensuring safety and dependability, “We are excited by the enormous potential of generative AI and by the challenge of pushing the limits of reliability of large language models.” Said Burger, “At atla, we are committed to creating safer AI systems that are designed to perform reliably in high-stakes situations. We’re thrilled to onboard our new high-conviction partners, led by Creandum, to execute our vision and drive the adoption of safer AI systems.”
Creandum’s investment perspective
Since 2003, Creandum has backed some of Europe’s most successful tech companies. Hanel Baveja, Principal at Creandum, highlighted the exceptional AI expertise and determination exhibited by Maurice and Roman. Creandum’s investment resonates with atla’s mission to pioneer reliable, safe, and trustworthy AI applications, particularly in sectors where precision is paramount.
Baveja, commented, “From our first interactions, we’ve been incredibly impressed by the ambition, relentless work ethic, and deep AI expertise from Maurice and Roman. We’re excited to join the atla team in their journey to build reliable, safe, and trusted AI applications for the sectors where this matters most.
