On February 2, the first part of the European AI Act came into effect. This means that some AI applications are now prohibited from use within Europe. As a business, it’s wise to be well-informed about what you can and cannot do with AI, especially if you rely heavily on it for your daily operations. In this article, we’ll tell you everything about the AI applications that have been banned in Europe since this month, so you’re fully up to date!
1. Artificial intelligence that manipulates or misleads
First, AI applications that manipulate or mislead people have been banned in Europe as of this month. Of course, this can be interpreted in many ways, because when does something actually constitute manipulation or deception? The European Court of Justice defines it as follows: “Systems that exploit people’s vulnerabilities, such as age or disability, to influence them.” Therefore, you may not use artificial intelligence that in any way exploits a person’s disability, or, for example, their advanced or conversely younger age.
2. AI applications that use social scoring are now prohibited
AI applications that use social scoring have also been strictly prohibited in Europe as of this month. You’ve probably heard of the term “social scoring.” It refers to assigning a specific rating to people based on their behavior—in other words, a social score. This means you can no longer use artificial intelligence that evaluates people based on their social behavior or certain personal characteristics. For example, if you use AI to scan job applications or resumes, you must ensure that this system does not evaluate applicants based on age, gender, or ethnicity. It has long been known that certain AI applications still have a strong tendency to discriminate, so be very careful with this!
3. Systems that exploit vulnerable groups
Unfortunately, artificial intelligence can also be used to exploit vulnerable groups in a “sophisticated” manner. Starting this month, the use of applications that do this in any way is also prohibited. Examples of vulnerable groups include the elderly, people with certain disabilities or impairments, people facing language barriers, or people in specific socioeconomic situations. AI applications specifically designed to exploit these types of groups may no longer be used.
4. Emotion recognition in the workplace and in education
Emotion recognition using AI in the workplace and in education is another form of AI that is now banned in Europe. But what exactly is emotion recognition, and how can an employer or teacher use it? In the workplace, it primarily involves the continuous monitoring of employees. This can be done in various ways, such as through eye-tracking, typing speed, or voice recognition. Another method we’re seeing more often these days is monitoring via wearables worn by employees. This is used to measure stress and fatigue levels, often to prevent accidents in the workplace.
Employers often claim to use these tools to promote the health and safety of their staff, but there is, of course, a significant downside. For employees, this leads to anxiety and a sense of being monitored. In addition, it can lead to an excessive workload and even burnout, emotional exhaustion, and reduced job satisfaction. For this reason, employers are no longer permitted to use AI for emotion recognition among staff.
5. Real-time biometric identification in public spaces
The final type of AI application that is now prohibited is AI that uses real-time biometrics to identify people in public spaces. This includes, for example, AI systems that identify people using real-time facial recognition. There are a few exceptions to this ban. Biometric identification in public spaces using artificial intelligence is permitted for the purpose of locating missing persons or preventing serious threats.
Ensure your organization has sufficient AI expertise
In addition to banning certain forms of AI, the European AI Act requires companies to ensure that their staff have sufficient knowledge of AI. This knowledge must be appropriate for their role within the organization and the context in which the company uses AI. It is important for your company to have this knowledge in-house, rather than relying on external experts. Failure to comply with the AI Act can result in very high fines of up to 35 million euros or seven percent of your company’s annual turnover. So make sure you stop using the AI applications that are now banned in the EU!