Transparency
Audiences should understand when AI materially shapes content, particularly where synthetic people, voices or stories could be mistaken for authentic testimony. Disclosure should reflect the risk of misleading the audience.
Business & innovation IWIB4AI session Needs review
The Impact of GenAI in the Communications Industry
Using generative AI in communications while protecting trust, creative rights, privacy and inclusive representation.
Draft – legal and case law references not yet verified
Summary
Generative AI can support communication strategy, content creation, storytelling and personalised audience engagement. This use case examines how those capabilities can advance diversity, equity, inclusion and belonging while introducing risks around misleading content, stereotypes, personal data and creative rights. Responsible adoption requires editorial judgement alongside technical capability.
The Problem
Faster production and more personalised messaging can improve audience engagement, but can also amplify stereotypes, obscure authorship and exploit audience information. The central challenge is preserving authentic, inclusive communication without sacrificing privacy or accountability.
Ethical Grounds
Inclusion
AI-assisted communication should represent diverse audiences without reducing them to stereotypes. DEIB goals require reviewing language, imagery and accessibility with input from the communities being addressed.
Privacy
Personalisation should use only necessary audience data and avoid unjustified inference of sensitive characteristics. Information collected for one communication purpose should not automatically become input for AI training or unrelated targeting.
Accountability
Communications teams remain responsible for the accuracy, rights clearance and effects of published material. Named reviewers and correction processes should prevent responsibility from being displaced onto the AI tool.
Legal Compliance
| Framework | Provision | Relevance |
|---|---|---|
| EU AI Act | Art. 50(2), (4) | Where applicable, these provisions distinguish provider duties to mark synthetic outputs from deployer disclosure duties for deepfakes and certain AI-generated public-interest texts, subject to specified exceptions. |
| GDPR | Arts. 5, 6, 9, 13–14 | Personalised communications involving personal data require a lawful basis, purpose limitation, data minimisation and appropriate information, with additional restrictions where special-category data are processed. |
| EU copyright law | Directive 2001/29/EC, Arts. 2–3 | Using protected material in AI-assisted workflows or publishing outputs that reproduce protected expression may engage reproduction and communication-to-the-public rights, requiring authorisation unless an exception applies. |
| EU consumer law | Directive 2005/29/EC, Arts. 6–7 | Consumer-facing AI-generated marketing must not mislead through false claims or material omissions, including misleading presentation of endorsements or commercial intent. |
Sources
Court & Regulator Decisions
Meta Platforms v Bundeskartellamt (opens in a new tab)
The decision clarifies legal constraints on combining audience data across services for personalised advertising, directly informing privacy-conscious AI communication strategies.
Infopaq International v Danske Dagblades Forening (opens in a new tab)
Protection can extend to short original text extracts, making source-material review important when generating communications from existing creative works.
Key Takeaways
- Assign human editorial responsibility for accuracy, rights clearance and corrections.
- Set disclosure rules that address audience expectations and applicable AI transparency duties.
- Test generated language and imagery for stereotypes, accessibility and inclusive representation.
- Minimise personal data used for personalisation and scrutinise sensitive audience profiling.
Related Use Cases
From the Session
Use Case #6 – The Impact of GenAI in the Communications Industry
Speakers
?mm&v=E3BC31A753E04B4FD40FE5514367C881A8CD2135){kind=avatar}
