Nvidia, Microsoft, Meta, IBM and several dozen other organisations have appealed to the US authorities not to restrict the development of open-source artificial intelligence models too soon. The companies argue that excessive regulation could stifle competition and drive innovation away from the United States.
Open-source models can be run in organisations’ own data centres, tailored to their specific needs and analysed for security. For businesses, this means greater control over data and reduced reliance on a handful of proprietary system providers. It may also reduce the costs of implementing AI, particularly in large companies, public administration and regulated sectors.
However, open models do carry risks. Once they are made publicly available, the creator has limited control over how they will be modified and used. This includes, amongst other things, cyber-attacks, disinformation and the copying of technology. The signatories of the letter believe, however, that these issues should be addressed through precise regulations rather than broad bans.
For Poland and the CEE region, the development of open-source AI could present a significant opportunity. Local companies can build their own solutions without incurring high fees for access to closed platforms. It will also become easier to create models tailored to the region’s languages and needs.
The most likely scenario is the development of a hybrid market. Companies will use both proprietary services and open-source models deployed locally. The deciding factors will be not only the capabilities of the models, but also infrastructure costs, access to specialists and security standards.
