AI solutions provider 2021.AI was in the AI game long before it became cool. Now, it’s inundated by requests to help companies implement AI responsibly.
For experienced AI companies like 2021.AI, ChatGPT changed everything. Business suddenly boomed.
“We were very happy about ChatGPT for three reasons,” says Mikael Munck, founder and CEO of 2021.AI, an AI company whose GRACE Governance Platform ringfences AI models so they can operate compliantly. “One—our families finally understood what we had been doing. Two—everyone using ChatGPT intuitively realised they needed a governance layer or solution. And, three—all our investment in the GRACE platform now paid off. Being model-agnostic, it integrates nicely with ChatGPT or any other AI model.”
Before starting his AI company, Munck worked at a major Danish bank that was one of the first to implement machine learning models. Despite these models’ efficiency, they were highly resource-intensive, complex, and lacked the community code support that exists today. Europe also lacked the talent for AI, and the bank had to source people from New York and London.
“Competition has intensified in the Nordic fintech sector over the last few years,” says Munck. “For fintech companies to maintain an edge, they’ll have to develop solutions that are technically more difficult for competitors to replicate. AI and LLMs will play a significant role in helping them do this.”
Mikael Munck, Founder and CEO of 2021.AI.
After leaving the bank, Munck wondered if there was a less-resource intensive way to develop such models. That’s when he met Danny Lang, a machine learning expert in the USA who had engineered precisely these types of platforms for Amazon and Uber, and provided valuable guidance.
AI Governance tech ahead of its time
After building for several years, 2021.AI became engaged with the EU three and a half years ago to help develop guidelines for building trustworthy AI.
“We recognised that, globally, there would ultimately be many regulations, criteria, and privacy needs around AI, both across jurisdictions and market sectors,” says Munck.
The company saw that its AI platform needed to do more than just support data scientists and organisations to increase their efficiency in developing and operating AI models. It also needed to deliver a full Governance, Risk, and Compliance (GRC) engine within its platform to ensure compliance with existing and future AI regulations.
Recommended: Explore the Future of Finance at Nordic Fintech Week
They developed GRACE, which ringfences a company’s AI and machine-learning models so companies can (a) efficiently develop, deploy and operate models at scale, and (b) provide such comprehensive GRC support that even the most regulated industries, such as healthcare and finance, would have all requirements covered.
“All the models currently in use in the banking sector—like the models used for credit scoring—must provide full documentation and transparency to ensure fairness,” Munck says. “This is precisely what GRACE provides.”
Any organisation faces a tough decision when it comes to implementing the latest breakthroughs in AI—Large Language Models (LLMs). Should they implement current LLM solutions and risk regulatory violations, or ban it outright and miss out on the phenomenal innovation and growth opportunity these models offer?
“If the healthcare sector can implement LLMs responsibly, any sector can.”
GRACE offers clients a complete suite of GRC solutions, including support for in-house LLMs.
GRACE’s LLM solution is currently deployed at Rigshospitalet, a highly specialised medical centre in Copenhagen. In this deployment, an LLM model trained exclusively on the hospital’s documentation will offer assistance to thyroid cancer patients and their families, even after hours.
Recommended: Explore the Future of Finance at Nordic Fintech Week
Coupled with GRACE, the model strips all Personally Identifiable Information and answers pre-operational questions for patients. If it encounters a question it’s forbidden from answering, it directs the query to a doctor, and it’s been specifically instructed never to offer medical advice.
“If the medical sector can compliantly implement an LLM using GRACE LLM Governance, then any sector can—even the financial sector,” says Munck.
For fintechs in the Nordics, AI and LLMs might be the make-break of success or failure, considering the competitive landscape.”Competition has intensified in the Nordic fintech sector over the last few years,” says Munck. “For fintech companies to maintain an edge, they’ll have to develop solutions that are technically more difficult for competitors to replicate. AI and LLMs will play a significant role in helping them do this.”
