The EU Passes its AI Act: Employers and Schools Should be Preparing for More Regulation Ahead

On March 13, 2024, a significant majority of the European Parliament voted to approve the EU Artificial Intelligence Act (the “AI Act”) .

The adopted full text of the AI Act can be found here.

This major step sets a course for a final version of the AI Act to be adopted in the Summer of 2024.  

Here are the key points organizations should be taking note of right now:

• Extraterritorial Impact outside the EU: Like the European General Data Protection Regulation, the AI Act will apply to organizations based in the EU, but will also include extraterritorial scope impacting companies not based in the EU.

• Banned Uses of AI: Certain applications of AI will be banned for giving rise to what the AI Act deems as unacceptable levels of risk. Banned uses include certain biometric categorization systems, data scraping public records for use in face recognition, and uses of emotional recognition systems in schools and workplaces.

• Managing High-Risk Uses of AI: The AI Act does not prohibit deemed high risk uses of AI, but requires such uses to satisfy accountability and risk management requirements that include human oversight and processes to minimize error and bias in data.

  
  

  Notably, high-risk uses of AI include the use of AI to affect decisions on admission to educational institutions, and to make employee recruitment decisions and other employment-related decisions. 

• Rights to Obtain an Explanation of AI Uses: The AI Act includes a right for individuals affected by a decision taken on the basis of high-risk uses of AI to obtain “clear and meaningful explanations” of the role of the AI system in the decision-making process.

  
  

With the official adoption of a final version of the AI Act now just around the corner, similar legislation can be expected in other jurisdictions, including Canada.

What does this mean for employers, schools, and other institutions?

We remain in the early days of AI adoption. For organizations, this means there’s no time like the present to take stock and avoid errors and oversights that can become costly down the line if not addressed at the earliest opportunity.  

The regulatory framework of the AI Act provides helpful guidance in this respect; and highlights the following as key considerations for organizations adopting AI uses in their processes:

1. Start with a risk assessment, considering the implications of decisions made by AI systems and the harm that could result from errors in these systems.

   
   

2. Establish boundaries based on risk, defining at the outset what uses of AI give rise to unacceptable levels of risk.

   
   

3. Maintain safeguards in the form of human oversight and scrutiny over the data used to train AI systems.

   
   

4. Establish transparency and accountability through processes that ensure affected individuals are aware of the organization’s AI usage and have an opportunity to call possible errors to the organization’s attention.

   
   

5. Stay tuned, as this is an area where technology and the law will likely develop quickly in the months ahead.

   
   

Contact inquiries@theprocesslegal.com to book a consultation on building a sustainable AI strategy.