Finnish companies generate innovation with AI

Despite some individuals’ hesitation to embrace the transformative power of generative AI, experts universally acknowledge its economic potential. A 2024 report from McKinsey estimates that generative AI could add 13 billion euros to Finland’s GDP by 2045.

Fuelled by Finland’s robust digital infrastructure and skilled workforce, AI is expected to drive productivity across various sectors, including business, law, STEM (science, technology, engineering, and mathematics), customer service, healthcare and construction.

Addressing challenges and ethical concerns

While much has been written about the productivity boost and possibilities that generative AI offers, there are also potential pitfalls. Generative AI faces challenges with bias and fairness, the risk of creating misinformation, and ethical issues surrounding intellectual property.

“Gen-AI has no built-in quality control. You cannot treat it as traditional software, but rather you need to think of it as an unreliable freelancer,” said Dr. Ari Heljakka, founder and CEO of the Finnish startup Root Signals. “You have to be pedantic in instructing it and then check its work in seven different ways – and then check again tomorrow.”

Based in both Helsinki and Palo Alto, Root Signals focuses on advancing AI applications in a controlled and reliable way. The company recently secured 2.8 million US dollars to support its innovative approach: it uses a “large-language model-as-a-judge” framework to monitor and evaluate AI outputs, providing a level of oversight critical for industries with high regulatory standards.

“We make this scalable with metrics that are understandable and easy to maintain in production,” Heljakka added.

Generating prosperity for Europe

While Root Signals’ focus on refining AI oversight represents an excellent example of early-stage growth, Finland’s reputation as a serious player in AI has also attracted attention from global tech giants. AMD’s acquisition earlier this year of Helsinki-based Silo AI for 665 million US dollars highlights the nation’s growing influence in AI innovation.

Europe’s largest private AI lab, Silo AI has worked with prominent clients such as Allianz, Philips and Rolls-Royce to integrate AI seamlessly into their operations. AMD’s acquisition sought to enhance its enterprise AI solutions by incorporating Silo AI’s advanced models and platforms.

Yet with AMD headquartered in the US, Silo AI continues operations from its established Helsinki base, offering a unique perspective on the differing tech growth trajectories between the US and Europe.

Indeed, in the shadow of an EU report about the future of European competitiveness, Silo AI’s CEO, Peter Sarlin, spoke out about fostering a robust AI ecosystem across Europe.

“Europe's tech competitiveness hinges not on research projects or regulatory frameworks, but on our ability to build and scale digital companies that can effectively embed AI into their offerings,” he asserted.

Silo AI has been practising as its CEO preaches. In late 2023, the company led a consortium of Finnish researchers and technologists in unveiling Poro 34B, a large-language model (LLM) designed to work proficiently in Finnish and English. Trained on a vast dataset using the LUMI supercomputer – Europe’s most powerful – this initiative marked the first step towards developing a family of LLMs covering all 24 official EU languages.

“From a European perspective, it is also critical that models are designed from the outset to prioritise [multilingualism] and an equitable approach to all languages,” said Sampo Pyysalo, a member of the natural language processing research team at the University of Turku playing a prominent role in the development.

Securing brings peace of mind

Poro 34B’s emphasis on multilingualism aligns with other efforts across Finland to make generative AI solutions accessible beyond the tech sector, with security at the forefront. ConfidentialMind, an Espoo-based startup, recently raised 1.5 million euros in pre-seed funding to enable enterprise and government clients to deploy LLMs securely and efficiently within their own data centres.

Unlike cloud-based AI solutions, ConfidentialMind’s software offers cloud-like ease of deployment directly on premises, addressing data protection and management challenges crucial for many organisations.

“We are seeing more and more interest from companies and governments in deploying their own open source generative AI-based applications, which tap into their enterprise or sovereign data,” said CEO and co-founder Markku Räsänen.

“This funding round shows that it is possible to build leading AI companies out of the Nordics,” he added.