The European Union has made a precedent setting decision regarding AI Laws.

At the end of last year, the European Union (EU) finally unveiled[1] its much-anticipated AI act. Following 36 hours of intense negotiations, all 27 member states reached a consensus, signing an agreement that, upon implementation (see below), will encompass systems like ChatGPT, as well as address the usage of facial recognition.

The primary objective of these regulations is to promote the responsible use of AI within Europe, aiming to ensure that “AI systems used in the EU are safe, transparent, traceable, non-discriminatory, and environmentally friendly.”

To achieve this, the proposed laws categorize AI systems into four groups: Unacceptable risk, High risk, and Low/Limited risk, with a separate category for generative AI and LLMs. The level of risk will be determined based on the number of floating-point operations per second (FLOPS) required to train an AI model.

Examples in each category include:

  • Unacceptable risk: performing scoring based on social behaviour or personal characteristics, emotion recognition in the workplace and biometric categorisation to infer sensitive data, such as sexual orientation;
  • High risk: AI used in sensitive systems, such as welfare, employment, education, or transport; and
  • Limited risk: chat bots or deepfakes.

 

However, conversely, the Act will not apply to AI systems:

  • used exclusively for military or defence purposes;
  • used solely for the purpose of research and innovation; and/or
  • used by people for non-professional reasons.

 

Additionally, the rules empower consumers to file complaints against AI developers. If found culpable, a newly established AI office (still pending) will have the authority to levy fines, reaching up to 7% of a company’s turnover, or £35m, whichever is greater.

Before the agreement receives the final nod from all Members of the European Parliament (MEPs), two committees[2] of the EU’s Parliament will conduct votes. Subsequently, both the Parliament and Council must formally adopt the agreed-upon text for it to finally become EU law.

As per the provisional agreement, the Act is slated to take effect two years after its entry into force, with certain provisions becoming operational at a later stage. Work finalizing the intricate details of the new regulation is still underway, suggesting that the Act may likely not come into effect until 2026.

Watch this space!

 

[1] https://www.europarl.europa.eu/news/en/headlines/society/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence

[2] The internal market and civil liberties committees.

 

Why Scintilla? We’ll put you in control of your destiny by organising, optimising and monetising your intellectual property.