Ethical issues of AI in advertising scandals 2025 and how brands can avoid mistakes

Artificial intelligence has become an integral part of digital marketing: it helps create texts, analyze audience behavior, optimize advertising campaigns, and even select images. However, alongside the growing popularity of these technologies, the number of ethical issues is also rising. In 2025, discrimination cases, privacy violations, and the use of biased algorithms have come into the spotlight.

Amid tightening regulations in the EU and the US, it is increasingly important for brands to understand the risks of using AI and how to avoid them.

Why AI can be biased

AI is trained on large volumes of data, often collected from open sources. These data may contain systemic biases such as:

• gender and racial stereotypes,
• socio-economic and cultural distortions,
• commercial biases favoring certain markets and languages.

Even if AI developers did not intentionally embed bias, it appears at the dataset level. In advertising, this manifests especially strongly — since AI directly influences who sees what.

Scandal of the year: discrimination in a fashion brand campaign

At the beginning of 2025, American fashion brand NOVA Frame launched a large-scale advertising campaign created entirely using AI image models (Midjourney and proprietary generative systems). The campaign featured visuals of “models of the future” — stylish, glossy characters designed for digital-first social media ads.

Within days, users noticed that all models had light skin tones, European facial features, and a “standardized” look. Instagram and X feeds lacked diversity, with most “images” stereotypically white and slim.

After an investigation published by the American magazine Wired, the company faced a wave of criticism:

• users and activists accused the brand of “digital racism,”
• the hashtag #AIBiasFashion trended with user critiques,
• the brand began losing advertising contracts and received a request from the US Federal Trade Commission (FTC) to explain its AI use principles.

It later emerged that the training dataset used for generating visuals mainly consisted of images from European and American fashion magazines from 2000–2020. The AI simply “mirrored” the dominant image embedded in the dataset.

Regulatory response

EU: AI Act and brand responsibility
Since 2025, the EU has enforced the AI Act — the first comprehensive AI law. According to it:

• AI advertising systems are classified as “high risk” if they influence consumer behavior,
• companies must audit models for discrimination,
• AI-generated content must be labeled (e.g., “AI-generated content” in descriptions and captions).

US: FTC powers increase and “AI Bill of Rights” discussion
In the US in 2025, the FTC gained expanded authority to oversee AI in advertising. Congress is also debating the introduction of an “AI Bill of Rights,” which would require:

• users to be informed if content is AI-generated,
• companies to demonstrate non-discrimination,
• an option to opt out of personalized AI advertising.

AI can dramatically boost advertising effectiveness — but only when used responsibly. In 2025, it’s clear that the technological race for consumer attention must not come at the cost of ethics, inclusivity, and transparency. Brands should implement ethical practices before regulators mandate them.