EU AI Act: Banned AI Uses Explained

EU AI Act: Initial Compliance Deadline and Guidance for Developers

The European Union’s AI Act, a risk-based regulatory framework governing artificial intelligence, saw its first compliance deadline take effect a few days ago.

This initial phase focuses on banning specific AI applications deemed to present “unacceptable risk.”

Prohibited AI Practices

Among the prohibited uses are practices like social scoring systems that could result in discriminatory or disadvantageous treatment.

Furthermore, the Act prohibits the use of AI for manipulative purposes employing “subliminal techniques” designed to cause harm.

Commission Guidance for Compliance

Following this deadline, the EU Commission released guidance intended to assist developers in adhering to these initial stipulations of the AI Act.

This advice is particularly relevant for those developing and deploying artificial intelligence solutions within the EU region.

Penalties for Non-Compliance

Violations of the regulations concerning prohibited AI applications can incur substantial penalties.

These penalties can reach up to 7% of a company’s global annual revenue, or €35 million, whichever figure is higher.

Nature of the Guidance

The Commission emphasized that these guidelines are designed to promote consistent application of the AI Act throughout the European Union.

However, it’s important to note that the guidance is not legally binding; enforcement and interpretation will ultimately rest with regulators and the courts.

Focus on Safety and Ethics

The Commission stated that the guidelines offer both legal clarifications and practical examples to aid stakeholders in understanding and meeting the AI Act’s requirements.

This initiative demonstrates the Commission’s dedication to cultivating a secure and ethical environment for artificial intelligence.

Availability and Ongoing Implementation

The draft guidelines are currently available for download.

Formal adoption and implementation are pending completion of translations into all of the EU’s official languages.

Although the AI Act was formally enacted last year, its full implementation will occur in stages over the coming months and years.

Staggered Enforcement

Enforcement of even the prohibited use cases will be phased, as EU Member States have until August 2nd to designate the authorities responsible for overseeing the AI Act.

• Further compliance deadlines are scheduled for the future.
• Member States are responsible for designating oversight bodies.