Employers should 'commit to fully understanding the Act's nuances to avoid becoming complacent,' says expert
HR professionals are assured that they can remain compliant with their AI use for immigration systems despite the recent high-risk classification of a new framework from the European Union (EU).
The EU AI Act, enacted in August 2024, has classified the use of AI in immigration services as high-risk in its new four-tier system.
This means that companies using an AI tool to perform an immigration-related task will need to fulfil a set of duties around transparency and risk management, or prepare a case on why the specific use case does not constitute a high-risk application of AI.
But Centuro Global said in a new report that the regulation can still exempt an AI tool from the high-risk classification if it meets any of the following criteria:
- The AI system is intended to perform a narrow procedural task.
- The AI system improves the result of a previously completed human activity.
- The AI system is intended to detect decision-making patterns or deviations from prior decision-making patterns and is not meant to replace or influence the previously completed human assessment.
Zain Ali, co-founder and CEO of Centuro Global, said the clause gives professionals the permission to use AI systems in global mobility, noting that these tools may already be meeting one of the exemptions.
"Most of the AI tools that HR teams are particularly likely to use will meet at least one of these criteria," Ali said in a statement.
"Of course, remaining compliant is key, so when using an AI tool, it's vital to stick to the developer's instructions, apply human oversight measures, and report any issues to the provider."
EU AI Act risk classifications
The EU AI Act is the world's first major regulatory framework on artificial intelligence. It assigns applications of AI to four levels of risk, namely unacceptable, high, limited, and minimal.
Applications and systems classified under the unacceptable risk are banned, such as the government-run social scoring of the type used in China.
Those under high-risk applications are put under specific legal requirements. This includes systems used in the immigration processes as well as CV-scanning tools that rank job applicants.
Other applications that do not belong to the first two categories are left unregulated, according to the Act.
HR urged to continue using AI
According to Ali, businesses don't need to be afraid of the act and should continue exploring the use of AI in the immigration system.
"HR professionals are already using AI to streamline and optimise other areas of their work, and there is a significant opportunity to enjoy the same benefits when doing business abroad," he said.
He reminded, however, that employers should completely understand the new AI Act given its complexity.
"If you're using AI, commit to fully understanding the Act's nuances to avoid becoming complacent, as non-compliance penalties could be financially and reputationally ruinous."