This morning, two groups of lawmakers in the European Parliament voted on a provisional agreement on European law on artificial intelligence. But what impact will it have on the use of AI tools for all of us European citizens? Bernd GreifenederCTO of Dynatracegave us his opinion on the impact of theAI Act just approved by the European Union. Analyzing not only geopolitical risks, but also those economical for companies.
The European Union approves the AI Act, what does it mean?
Second Greifenederthe support expressed by legislators makes it almost inevitable that the law actually becomes lawopening a new chapter in the regulation ofArtificial Intelligence (AI). EU regulation appears to focus on reducing the geopolitical risks ofAIbut Greifeneder noted that it does not directly address commercial risks.
According to Greifeneder, many organizations may choose to use templates AI which fall into the category defined by the EU as “minimal or no risk”. However, he stressed that “most organizations will likely find themselves using what the EU calls “little or no risk” AI models. These models do not fall directly under the influence of the AI Act, but the EU is encouraging organizations to engage in voluntary codes of conduct to better manage risk.”
Greifeneder highlighted the importance of developing suitable codes of conduct, recognizing that not all AI is the same. He emphasized that business managers must classify theAI basically you have risk parametersconsidering the potential impact on revenues, reputation and stakeholder relationships.
We need to understand: “how AI makes decisionsi, whether it is transparent and which processes it has access and control to. They should identify whether its results are deterministic, derived from relevant and contextual data updated in real time, and therefore highly accurate, or whether AI draws conclusions from random and closed data, making her prone to errors and hallucinations.”
Furthermore, Greifeneder cautioned that without a clear classification framework delineating these characteristics, organizations risk using AI unsafely, even if they comply with the legislation. In short: the AI Act alone is not enough — you need to know the artificial intelligence you are using well.
