EU’s Regulations and the Future of AI
June 22, 2023
Artificial intelligence (AI) is advancing at a rapid pace, faster than we might imagine. However, It is not only fascinating but also, at least to some of us, unpredictable and scary. This is why the European Commission is currently contemplating what regulations and guidelines should be established for AI.
Will these new rules hinder AI development? And will they affect startups that utilise or develop AI technology?
High-risk areas of AI include, for example, infrastructures involving ethical considerations and safety components of AI products. However, the lack of clear regulations can also keep companies on edge, as they are uncertain about the guidelines. This uncertainty can impact the scale of investments companies are willing to make in AI.
Nevertheless, both Mikko Lehtimäki, Co-founder and Chief Data Scientist at Softlandia, a company leveraging AI, and Pontus Stråhlman, partner at Voima Ventures, which invests in AI startups, believe that regulations are ultimately beneficial.
“AI is a tool, just like many others. Of course, it is a challenging tool and can be misused, so having some ground rules is essential”, says Stråhlman. “The purpose of regulation is good, and it does not inherently hinder AI development.
If AI were left unregulated, surveillance technology could develop in a detrimental direction”.
Lehtimäki shares the same view. “Surveillance applications definitely need regulations”, he confirms. “The EU’s risk categorization is reasonable, however the regulations should be considered based on the applications that utilise AI, rather than solely on whether a technology is classified as AI or not”.
He also describes AI as “a vague concept” and believes it would be better to focus on rules based on the use cases rather than merely the presence of AI. This could be the biggest challenge when the EU starts implementing regulations.
How will regulations affect AI startups?
Voima Ventures has invested in several startups that leverage AI, such as Kuva Space and MVision. These companies employ AI in various ways, including data collection and medical applications. Stråhlman notes that regulations like GDPR have posed clear challenges in using medical data.
However, he has a clear message for startups: don’t be too afraid of AI regulation.
“Startups should focus more on whether their business ideas will still be relevant in 5-10 years when AI progresses in great leaps and bounds”, Stråhlman emphasises.
Again, both Stråhlman and Lehtimäki agree that excessive regulation of AI can be a hindrance. “I certainly don’t oppose privacy protection, but regulations related to it slow down the progress”, says Stråhlman.
Lehtimäki mentions that EU regulations have not yet affected their company’s operations, but the unclear nature of the rules makes it difficult to draw conclusions.
“It is possible that once the regulations come into force, we may have to register as a provider of large-scale AI models”, he notes. “This would entail additional costs and reduce our agility, and we might need to demand reports from our customers on the impact of using our applications”.
Even though the regulations have not been finalised yet, their implementation will most obviously increase the costs and slow down the innovation. “If utilising or offering AI technology becomes too bureaucratic, it will impede the operations of small companies”, explains Lehtimäki.
“AI will become the business of large corporations, meaning there will be fewer players innovating and developing AI and AI-powered services”.
The future of AI
Despite forthcoming regulations, Lehtimäki does not believe they will significantly slow down AI development: “Of course, if AI regulations elsewhere in the world are not as burdensome as in the EU, there is a risk of falling slightly behind global progress, but probably not significantly”.
Stråhlman and Lehtimäki underline that regardless of regulations, AI will be part of everyday life in Finland and globally in ten years. They do not anticipate an AI dystopia or a scenario where AI surpasses human intelligence.
“Human beings can create much more dangerous weapons that pose a greater risk than AI”, says Stråhlman.
Instead, he believes that AI will solve many societal challenges, such as labour shortages. However, the nature of work will undergo significant changes, presenting both opportunities and challenges. It is also important to remember that if AI is not subjected to any regulations, it will undoubtedly be used for nefarious purposes.