Dutch startup launches open-source language for reliable AI

With the launch of the open-source language Lemma, the Dutch startup LemmaBase is taking a concrete step to address the power of Big Tech and the risks of unpredictable AI. This new standard enables governments and companies to codify laws, regulations, and policies in a transparent and verifiable manner.

The startup is thus offering a direct response to the growing dependence on opaque algorithms and AI systems that are putting the autonomy of governments and businesses under pressure. According to LemmaBase, society is at risk if decision-making regarding financial legitimacy, benefits, permits, or loans is left to systems that cannot be monitored or controlled.

Making code understandable

Lemma solves this by applying the “rules as code” principle to AI. The new language makes it possible to formulate laws and regulations in a way that is easy for humans to understand and can be executed by computers with absolute certainty. It shows what logic is being applied, why a specific outcome is reached, and how rules change over time.

Banks and health insurers

Such guarantees are crucial in sectors where errors are not an option. In the financial sector, for example, an AI assistant can handle the interaction during a mortgage or loan application, while Lemma ensures that borrowing capacity is calculated according to applicable standards. The government can embed tax logic or benefit eligibility criteria in Lemma, making decisions explainable and eliminating arbitrariness. Municipalities can have complex zoning plans reviewed during permit applications to ensure faster and legally sound decisions, while health insurers and logistics service providers can use the language to guarantee error-free healthcare reimbursements and compliance with international customs regulations, respectively.

AI governance

The launch of Lemma aligns with a broader trend in which the focus is shifting from AI capabilities to AI governance. Whereas the technology has primarily been used for generative tasks in recent years, there is growing demand for systems that offer firm guarantees. Decoupling decision-making logic from operational systems gives organizations a way to innovate without losing control.