Universal Registration Document 2025

IROs/Score/ Time frame

Definition of IRO

Policies & ad hoc work

Major actions [ESRS 2 MDR-A-68 (a)&(b)&(c)] [S4-31 (a)&(c); S4-33 (a)&(b); S4-34]

Objective

Negative impacts

IRO 15

Entity Specific

Major

ST/MT/LT

Consumers’ personal data may be impacted if protection and security rules are not respected,

Janus – Values
Janus – Compliance & Ethics
Janus - Data Protection
Janus - Data Security
Cybersecurity risk mapping

Data security is one of the mandatory training courses for all employees.
Ad hoc governance’s in place for artificial intelligence tools to analyze and validate them before their use in the Company and for projects intended for clients.
Data security is managed by the GSO (Global Security Office) teams. All employees—upon joining the Groupe, and then every year—must complete training to refresh their knowledge of the issues and best practices in this area. The GSO is integrated into client projects from the outset, in order to anticipate technological challenges, in conjunction with the customer’s technical teams.

90% of employees trained every year

Negative impacts

IRO 16

Entity Specific

Major

ST/MT/LT

The use of AI could infringe end-users’ rights, in particular by creating discriminatory situations

Janus – Values
Janus – Compliance & Ethics
Janus - Data Protection
Janus - Data Security
Cybersecurity risk mapping

Epsilon, for example, has set up a whole process for reviewing proposals made by AI to ensure that the program will not harm consumers and end users. The latter may at any time decide to exit actions initiated by Epsilon ("opt-out")
Since 2023, Publicis GPT is in place within the Marcel.ai internal platform has been in place for employees to test these tools in a protected environment.
An employee training program is in place to support skills development. In 2025, all employees were trained in the Generative AI Ethics and Responsible Use module.

90% of employees trained

Opportunities

IRO 17

Entity Specific

High

ST/MT/LT

The use of Generative AI is a powerful tool to broaden the possible options, whether they are creative or in the dissemination of the right message at the right time to the right person, on a very large scale.

Janus – Values
Janus – Compliance & Ethics
Generative AI Acceptable Use Guidelines
AI Legal Guidelines.
Janus - Responsible Marketing

Employee training is the key lever for ensuring that everyone embraces these new tools, both in terms of innovation and understanding their limitations. The aim is to ensure that the use of AI will not harm consumers
A space dedicated to AI called PL.AI in Marcel allows employees to test many AI tools
Working with internal experts also allows for the sharing of constantly evolving knowledge

90% of employees trained each year

IROs: impacts - risks - opportunities

Score: Low, Medium, High, Major

Time frame: ST: short-term; MT: medium term; LT: long-term

4.3.11.2 Impacts on consumers and end-users

Communication has always supported the development of companies, enabling them to promote their products and services to as many people as possible. In a highly fragmented and rapidly changing digital world, Publicis Groupe’s BtoB (Business to Business) strategy is to help its clients deploy effective marketing and communication actions through large-scale personalization of messages. The first step is to know consumers, their expectations and their needs. For decades, the Groupe’s agencies have relied on studies conducted directly or indirectly with the help of specialized and independent third parties, providing accurate factual information.

Three negative impacts were analyzed:

The messages of a campaign or the functioning of an application, for example, could cause environmental or societal harmif they do not respect ecological transition messages or if they could hurt a certain group of people. The teams are made aware of the societal impact of their actions and messages, in order to anticipate possible reactions. The objective of the Groupe’s business lines is to respect the many sensitivities, preserve the brand, product and service of clients, and nurture the relationship of trust with consumers.
The use of good professional practices is the key to preventing a campaign from being poorly received and harming consumers on an environmental or social level. The use of artificial intelligence (AI) raises legitimate concerns among the general public. With the arrival of generative AI came training courses for all employees to familiarize themselves with the various tools, especially to learn how to use them responsibly, and the limits on use to be respected. [ESRS S4 SBM-3-11]
In a digital world dominated by data, it is essential to ensure that datasets are used responsibly, and that AI solutions are trained to avoid harmful biases. Similarly, consumers must be given the opportunity at any time to access their data and/or to exercise their right to be forgotten. [ESRS 2 MDR-P]

Universal Registration Document 2025

Board of directors

4.3.11.2 Impacts on consumers and end-users