Simply ‘using’ generative artificial intelligence ‘properly’ is no longer enough. From 2 August, the obligations set out in Article 50 of the AI Act – the Regulation through which the European Parliament sought to ensure that artificial intelligence systems used in Member States are safe, transparent, traceable, non-discriminatory and environmentally friendly – which also introduces transparency obligations for those using generative systems for professional purposes, producing content that informs the public on matters of general interest. Not only must interactions with users be ‘labelled’ so that it is clear they are conversing with a machine rather than a person, but artificially generated or manipulated content (images, audio and video) – including deepfakes – and informative texts must also clearly indicate that they are the result of AI.
Meanwhile, the European Commission has published the final version of the “Code of Good Practice on the Transparency of AI-Generated Content”, which was drafted under the guidance of six independent experts in collaboration with over 180 stakeholders from the sector, including businesses, academics and civil society organisations. The aim of ensuring transparency stems from the need to help people recognise AI-generated content, thereby reducing the risks of deception and manipulation. The Commission has also created a set of icons to label content generated and manipulated by artificial intelligence. Texts generated by artificial intelligence that undergo a process of human review or editorial oversight, and where a natural or legal person is editorially responsible for the publication of the content, are exempt from the labelling requirements.
Faced with this necessary revolution, one has to wonder what will change for businesses, given the rapid adoption of artificial intelligence not only in processes but also in marketing and communication activities. Many companies are already using generative tools to be faster, more creative and more efficient. From text to images, from synthetic voices to personalised campaigns, right through to social media.
With the AI Act coming into force, rather than focusing on what can or cannot be done with artificial intelligence, attention must be paid to how it is done in a transparent and responsible manner. The legislation has a very clear objective: those consuming content must not be led to believe that it is entirely authentic or produced entirely by one or more people, when in fact it has been produced or significantly altered using artificial systems. Consequently, the use of artificial intelligence in communication can no longer be treated as a ‘technical’ or ‘creative’ matter, but will soon become a question of governance, compliance and even reputation. It will therefore be necessary – indeed, legally mandatory – to correctly disclose the use of artificial intelligence, by establishing business processes capable of ensuring transparency.
Copy and images for promotional campaigns, automated adaptations of messages, social media content, newsletters… thanks to generative tools, content can not only be produced more quickly but is often more effective, persuasive and even emotionally engaging. However, the people for whom these messages are intended (customers, users, stakeholders) must be able to recognise their artificial nature.
The impact of the AI Act on businesses – particularly on marketing and communications teams – will, in practice, involve a sort of review of their workflows, mapping the AI tools in use, distinguishing between the different types of content produced, defining internal responsibilities and putting in place checks prior to communication. This will involve a shift from a ‘spontaneous’ use of artificial intelligence to a governed and traceable approach. This will entail clear rules and safeguards in contracts with suppliers, agencies, content creators and technology partners to define responsibilities regarding content traceability and compliance with transparency obligations.
Just another obligation? Beyond the constraints, the new European legislation can offer businesses the opportunity to use transparency as a distinguishing feature and, therefore, as a competitive advantage. In a communications ecosystem now saturated with short-form content, deepfakes and creative automation, transparency, awareness and accountability are exceptions that can translate into trustworthiness.
