AI in advertising: Key aspects for advertisers in Sweden

Executive Summary – Key Takeaways

• General marketing compliance: All AI-generated content used in marketing must adhere to the requirements set out in applicable marketing regulation. Businesses bear full responsibility for their use of AI in marketing, including any misleading information or exaggerated claims about a product or organization. 
• Governance and review: Use of AI and incorporation of AI-generated content triggers various obligations and liabilities. Clear governance and review procedures should be established before publishing AI-generated content.
• Use of IP: AI-generated material may inadvertently reproduce protected materials such as trademarks or copyrighted works, which may lead to infringements in third party rights. Internal guidelines and manual review are essential to mitigate the risk of potential infringement.
• Liability for AI-generated content: Incorrect or misleading AI-generated content may constitute unfair marketing. This applies to chatbot responses, product descriptions, advertisements, and any other content. Advertisers bear full responsibility for ensuring accuracy, regardless of whether content is produced in-house or by a service provider, and whether it is produced by humans or AI.
• Identification of AI-generated content: As of 2 August 2026, the EU AI Act will require advertisers to disclose when certain content is AI-generated, particularly for deepfakes and AI-generated text on matters of public interest. The ICC’s guidance on responsible AI in marketing may also call for disclosure where needed to prevent consumer deception.
• Data privacy: AI-driven advertising applications that process personal data must comply with applicable privacy and data protection requirements, such as the obligations of the GDPR.
• AI literacy: Advertisers using AI tools usually qualify as deployers under the EU AI Act and must ensure a sufficient level of AI literacy among staff.

Background

The emergence of generative AI tools has transformed how advertisers create and distribute marketing content, raising important questions about compliance with key regulatory frameworks, including the EU AI Act[1] and national marketing legislation. This article aims to highlight some of the key general points that advertisers should consider when using AI tools for marketing purposes.

When AI is used in advertising, advertisers must ensure compliance with applicable marketing legislation, notably the Swedish Marketing Practices Act[2]. As Sweden’s primary marketing legislation, this Act governs unfair marketing practices in both B2C and B2B contexts and requires all marketing to adhere to good marketing practice.[3] In assessing what constitutes good marketing practice, guidance is partly drawn from the ICC Code[4], which, although not legally binding, is widely recognised as an industry standard and is regularly referred to by Swedish courts as an expression of good marketing practice. Notably, the ICC Code expressly provides that marketers who use algorithms or other AI-based solutions remain responsible for the results these produce.[5]

Key provisions of the Swedish Marketing Practices Act include:

1. Unfair marketing practices: Marketing that violates good marketing practice is considered unfair if it notably affects, or is likely to affect, the recipient's ability to make a well-informed business decision.[6]
2. Advertising identification: All commercial communication must immediately and clearly appear as advertising and identify the sender.[7]
3. Prohibition against misleading advertising: Statements must be true, current, and verifiable. It is prohibited to provide false information or provide material information in a way that misleads the target audience.[8]

The abovementioned rules are technology-neutral, meaning that the same rules apply regardless of how content is created or distributed. The use of AI tools does not exempt advertisers from these requirements. Advertisers remain fully responsible for ensuring that all marketing materials, whether generated by AI or through traditional means, adhere to the Swedish Marketing Practices Act and other applicable legislation. The liability remains the same regardless of whether the materials are produced in-house or by a third-party service provider, such as a creative agency, who may also be liable as a contributor under the Swedish Marketing Practices Act.[9]

Non-compliance with the Swedish Marketing Practices Act may result in enforcement action by the Swedish Consumer Agency (Sw: Konsumentverket) or proceedings before the Patent and Market Court (Sw: Patent- och marknadsdomstolen). Sanctions may include injunctions, prohibitions and market disruption fees. In cases of misleading or aggressive marketing, affected parties may also seek damages. These consequences apply equally to AI-generated marketing content.

In addition to marketing law, advertisers using AI tools should also consider their obligations under the GDPR[10]. AI-driven applications, such as chatbots or personalised advertising systems, often involve the processing of personal data. Advertisers should ensure compliance with applicable GDPR requirements, including establishing a valid legal basis for processing, providing transparency information to users, and implementing appropriate safeguards where third-party AI providers are involved or where data is transferred outside the EEA.

Liability for AI-generated content

In a recent judgment dated 1 April 2026[11], the Patent and Market Court ruled that incorrect responses from an AI chatbot may constitute misleading marketing under the Swedish Marketing Practices Act.

The case concerned a mobile operator that used an AI-powered chatbot for customer service. When a consumer asked for a customer service phone number, the chatbot incorrectly stated that the company did not offer telephone support. The Court held that this response was misleading and constituted unfair marketing.

A key takeaway from the ruling is that responses from chatbots and other automated response systems qualify as marketing within the meaning of the Swedish Marketing Practices Act, regardless of whether the responses are sales-driven. This means that the Act’s requirements apply to automated responses generated by such systems.

The ruling has broader implications beyond this specific case. First, the principle likely extends to other types of incorrect chatbot responses. If a chatbot provides inaccurate product information, incorrect pricing, or false promotional claims, this may also constitute unfair marketing. Second, the principle is likely not limited to chatbots. Automated telephone response systems and auto-reply emails may be subject to the same reasoning. For example, if an automated voice system informs a consumer that complaints are handled exclusively through the company’s website, when other channels are in fact available, this could also constitute unfair marketing.

A practical consequence of this case is that it is no longer sufficient for advertisers to review only advertising campaigns and promotional materials for legal compliance. Automated systems that respond to customer enquiries must meet the same legal standards as traditional marketing communications.

Further, the Swedish Advertising Ombudsman (Sw: Reklamombudsmannen) has reached a similar conclusion regarding AI-generated marketing content. For reference, the Swedish Advertising Ombudsman applies the ICC-Code in its supervision. In a decision from 13 August 2025[12], an education company was found to have published misleading information about career paths to becoming a nurse. The company acknowledged that the incorrect information had been generated by an AI tool during an internal experiment. The Advertising Ombudsman held that the use of AI technology did not excuse the misleading nature of the content. While decisions from the Swedish Advertising Ombudsman are not legally binding, they still serve as guidelines for what can be considered misleading marketing.

The key lessons from these cases are clear:

• AI is not an excuse for unfair or misleading marketing.
• Advertisers remain fully responsible for the accuracy of their content, regardless of whether it was produced by a human or generated by an AI tool.
• AI-generated material must be reviewed with the same care as manually produced content.

Risk of IP infringement

AI-generated material may inadvertently infringe third-party intellectual property rights, including trademarks and copyrights. Since AI models are trained on vast datasets potentially containing protected content, there is an inherent risk that outputs may reproduce or closely imitate protected works. Advertisers using AI for marketing should therefore implement safeguards to identify and mitigate these risks.

With respect to trademarks, the primary risk arises when AI models recreate elements that could be confused with registered trademarks (e.g., an AI image generator producing a logo closely resembling a well-known brand). This may occur visually (e.g., logos or brand imagery) or textually (e.g., brand names or slogans), either because the prompt contains such elements or because the model’s training data carries signals from registered trademarks.

With respect to copyright, the risk is equally significant. When AI-generated text, images, or music are used in advertising, there is a considerable risk of reproducing or closely imitating protected works without this being apparent to the advertiser. Using such material without authorisation can lead to disputes and claims for infringement from rightsholders.

To manage these risks, advertisers should develop internal guidelines prohibiting employees from using others’ registered trademarks or protected works in prompts. AI-generated material should also be manually reviewed before publication, with particular focus on the risk of confusion with registered trademarks as well as potential copyright infringement.

Identification of AI-generated content and AI literacy

Article 50 of the EU AI Act will enter into force on 2 August 2026, introducing transparency requirements for both providers and deployers of AI systems. A deployer is any person or organisation that uses an AI system in a professional capacity (as opposed to purely personal use). Advertisers using AI tools to create marketing materials will therefore be subject to these rules, and will primarily face the following two obligations:

1. Deepfakes: When using AI to create realistic images, audio, or video of people, places, or events (commonly known as “deepfakes”), advertisers must clearly disclose that the content is AI-generated.[13] This applies whenever the content could reasonably be mistaken for real footage or recordings. However, if the deepfake is part of an obviously artistic, creative, satirical, or similar work, the disclosure requirement is more limited. In such cases, advertisers need only indicate the presence of AI-generated content in a way that does not interfere with the display or enjoyment of the work.
2. Text informing the public on matters of public interest: When using AI to generate text that is published to inform the public on matters of public interest, advertisers must disclose that it is AI-generated.[14] However, this does not apply if the text has been reviewed and edited by a person who takes editorial responsibility for the content. It is worth noting that this text disclosure requirement only applies to publications on matters of public interest. AI-generated marketing that does not address matters of public interest falls outside this specific obligation.

Any required disclosure must be presented clearly and prominently to the audience, at the latest when they are first exposed to the content.[15]

Beyond the provisions of the EU AI Act, the ICC has issued a guide on responsible AI in marketing[16] which provides further guidance on when disclosure of AI use may be warranted. According to the guide, the fact that generative AI output or AI systems were used to create advertising or marketing communications does not, in itself, require disclosure. However, disclosure may be necessary where it is needed to prevent consumer deception or to counter a potentially misleading impression created by the content. It should also be noted that a disclosure that AI has been used will not, in itself, cure claims that are false or misleading for other reasons.

To prepare for these requirements set out in the EU AI Act, advertisers should review how they use AI in content creation and establish clear processes for labelling AI-generated material where required. The European Commission is currently developing practical guidance on how to comply with these rules, which is expected to be available before August 2026.[17]

Further, Article 4 of the EU AI Act, applicable since 2 February 2025, requires deployers of AI systems (including advertisers using AI tools) to ensure, to their best extent, a sufficient level of AI literacy among staff and other persons operating or using AI systems on their behalf, taking into account their technical knowledge, experience, education and training, and the context of use. In practice, advertisers should implement internal training and awareness measures so that staff using AI tools for marketing have an adequate understanding of their capabilities, limitations and risks.

Concluding remarks

AI offers significant opportunities for advertisers, but it also introduces new legal risks. Compliance requires the advertiser to take a proactive approach, including:

• Establishing clear internal guidelines.
• Implementing review processes for AI-generated content.
• Confirming that AI-generated content does not infringe third-party intellectual property rights.
• Regularly monitoring chatbots and other automated customer-facing systems to ensure they provide accurate information.
• Ensuring a sufficient level of AI literacy among staff using AI tools.
• Keeping up to date and informed as new rules and guidelines enter into force.

Advertisers who implement these safeguards now will be better positioned to utilize the benefits of AI while avoiding regulatory risks. Further, the application of marketing rules to AI-generated content is still developing. While some guidance can be drawn from recent rulings and decisions, it remains to be seen how regulatory authorities and courts will interpret and enforce these rules in practice. Case law will continue to shape the legal landscape in the years to come.

References

[1] Regulation (EU) 2024/1689 (AI Act).

[2] The Swedish Marketing Practices Act (2008:486).

[3] Section 5, The Swedish Marketing Practices Act (2008:486).

[4] The International Chamber of Commerce’s Advertising and Marketing Communications Code.

[5] Page 14, The International Chamber of Commerce’s Advertising and Marketing Communications Code.

[6] Section 6, The Swedish Marketing Practices Act (2008:486).

[7] Section 9, The Swedish Marketing Practices Act (2008:486).

[8] Section 10, The Swedish Marketing Practices Act (2008:486).

[9] Section 23, The Swedish Marketing Practices Act (2008:486).

[10] Regulation (EU) 2016/679 (General Data Protection Regulation).

[11] PMT 13029-25.

[12] Matter 2503-36, the Swedish Advertising Ombudsman.

[13] Article 50.4 first paragraph, Regulation (EU) 2024/1689 (AI Act).

[14] Article 50.4 second paragraph, Regulation (EU) 2024/1689 (AI Act).

[15] Article 50.5, Regulation (EU) 2024/1689 (AI Act).

[16] The International Chamber of Commerce’s Guide for responsible AI in marketing.

[17] European Commission, “Commission publishes second draft Code of Practice on marking and labelling of AI-generated content” (available at https://digital-strategy.ec.europa.eu/en/library/commission-publishes-second-draft-code-practice-marking-and-labelling-ai-generated-content, last accessed 6 May 2026).