Artificial intelligence (‘AI’) is revolutionising various industries and in today’s consumer-driven world, ensuring the safety of products is of paramount importance, more so when considering the innovative products being placed on the market.

At the core of the AI Act [1] are requirements that providers, manufacturers, importers, and distributors of AI systems must meet before making an AI system available for use or distribution on the Union market. These requirements cover AI systems that are, or that are safety components of, products such as toys, lifts, radio equipment, and medical devices which are classified as “high-risk” – that is AI systems that pose a high risk of harm to the health and safety or the fundamental rights of persons.

While the introduction of the AI Act is an important step forward to ensuring that product safety is upheld in today’s digital world, it should not detract attention from the General Product Safety Regulation (the ‘Regulation’) [2] which, as from the 13 December 2024, will replace the current General Product Safety Directive (the ‘Directive’) [3] (as well as the Food Imitating Product Directive).

The AI Act and the Regulation will be briefly considered in turn below.

Product Safety under the AI Act 

Providers of high-risk AI systems that fall within the remit of the AI Act have the responsibility to comply with several requirements, some of which are akin to those found in product safety specific legislation. Such requirements include:

– indicating on the AI system, or its packaging where the former is not possible, their name, registered trade marks or trade names, and the address where they can be contacted;
– drawing up the EU declaration of conformity; and
– affixing a CE marking to the high-risk AI system or its packaging.

Providers must also draw up technical documentation of the high-risk AI system before placing the same on the market or putting it into service. The technical documentation, which must be drawn up in such a manner as to provide the relevant national competent authorities with the necessary information to assess the compliance of the AI system with the requirements of the AI Act, should be retained for a period of ten years.

Where the high-risk AI system relating to a product covered by Union harmonisation legalisation – including, to name a few, the Directives on the safety of toys; on medical devices; and relating to lifts and safety components for lifts – is placed on the market, the drawing up of a single set of technical documentation will suffice insofar as this documentation contains the information laid out in the AI Act as well as the respective Union harmonisation legislation.

Importers of high-risk AI systems have similar responsibilities, including ensuring that the provider has drawn up the requisite technical documentation before placing such systems on the market and that the system bears the CE marking. Where the provider is established in a third country, the importer must also ensure that an authorised representative has been appointed. In the case of distributors, these must ensure that the high-risk AI system bears the CE marking and that it is accompanied by a copy of the EU declaration of conformity.

General Product Safety

The Regulation reproduces a number of provisions present in the Directive – its aim is to update the Directive in the light of new technologies that have emerged. In terms of the Regulation, manufacturers of products must ensure that other persons in the supply chain are kept informed, in a timely manner, of any safety issues that they have identified with respect to a product. Manufacturers must also make publicly available communication channels, taking into account accessibility needs for persons with disabilities, to enable consumers to submit complaints and to inform manufacturers of any safety issues they have experienced with a product.

Providers of online marketplaces must register with the Safety Gate Portal, on which they must indicate the information concerning their single contact point. Furthermore, in terms of the Regulation, the European Commission is to develop, modernise, and maintain the Safety Gate Rapid Alert System for the exchange of information on corrective measures concerning dangerous products. The local national authority – the MCCAA – will then be required to notify through the Safety Gate Rapid Alert System any corrective measures taken by it or by economic operators in relation to the Regulation.

Over the coming weeks, manufacturers and online marketplaces must analyse their current product safety practices to ensure they are compliant with the new Regulation.

Concluding Remarks

Both the AI Act and the Regulation are welcomed pieces of legislation in the field of product safety – while the Regulation introduces several new obligations on manufacturers and online marketplaces, the AI Act acts as a much-needed step forward towards regulating the field of AI.

Business will have a few months over the course of which they must ensure that their operations are compliant with the AI Act. However, with the Regulation coming into force in December, manufacturers and online marketplaces must analyse their current product safety practices to ensure they are compliant with the same over the coming weeks.

This is the sixth in a series of articles exploring the effect of the AI Act across various sectors and industries.