AI in Advertising in 2026: Examples, Reactions and Disclosure Rules

AI in advertising before the generative wave

Machine learning has shaped advertising for well over a decade, mostly out of sight. Search engines and social platforms use it to decide which ad you see, how much the advertiser pays and which headline variant wins. The arrival of tools that generate images, video and voice changed what AI in advertising meant to the public.

The change was fast. By 2024, brands could produce a finished-looking spot without a camera crew, and several tried. The early results attracted attention mainly because they were made by software, not because the ideas were strong. That pattern, novelty first and craft second, shaped how audiences responded to almost every example that followed.

Early examples: a toy store and a holiday truck

In 2024 Toys “R” Us released a short brand film made with OpenAI’s Sora video generator. It showed the company’s founder as a boy dreaming of Geoffrey the Giraffe. Reviewers pointed out unnatural movement, odd visual artefacts and a pair of glasses that changed shape. The film was promoted as a first, and that was about all it achieved.

Coca-Cola went further in November 2024, releasing three short AI-made versions of “Holidays Are Coming,” its 1995 Christmas spot with the glowing red trucks. The backlash on social media was immediate, with artists including Alex Hirsch, creator of Gravity Falls, criticizing the work. The company defended it and released another AI-made holiday ad in November 2025 anyway.

2025: cheap spots, fake models and a mixed verdict

In June 2025 the prediction market Kalshi aired a chaotic ad during the NBA Finals, built with Google’s Veo 3 and ChatGPT by a self-described AI filmmaker. Its production cost was reported at around $2,000. Viewers called it strange, which may have been the point, and marketers saw it as proof that broadcast-quality footage no longer needed a big budget.

Fashion produced the sharper reaction. In August 2025 Guess ran an advertisement in Vogue featuring models that were generated by a small studio rather than photographed. Critics online argued that the images set impossible standards and removed paid work. The episode showed that AI in advertising draws the harshest criticism when it replaces people the audience can see.

There was also a quieter trend: brands using generative tools for storyboards, mood films and dozens of social cut-downs, then finishing with a conventional shoot. This hybrid approach rarely makes headlines. It is, as of 2026, probably the most common way AI enters a campaign, though few companies publish numbers on it.

The ad platforms pushed in the same direction. Google and Meta both added generative features to their self-serve ad tools, suggesting backgrounds, image variations and headline copy inside the same screens where budgets are set. For a small business, that is where AI in advertising actually shows up: a button that offers ten versions of a product photo.

AI in advertising at Super Bowl LX in 2026

Super Bowl LX in February 2026 became a showcase. Industry trackers counted roughly a fifth to a quarter of the spots featuring AI in some form, either selling AI products or using generative tools. The Super Bowl commercials page covers the full night, from Budweiser’s winning Clydesdale to the polar bears.

Svedka’s vodka ad was billed as the first Super Bowl spot made primarily with generative AI. Its dancing robots were widely mocked for their uncanny faces, and the ad finished near the bottom of USA TODAY’s Ad Meter. Svedka said it expected a polarizing response. Anthropic, an AI company, ran a spot joking about plans to put ads inside chatbots.

Disclosure rules in 2026: EU AI Act Article 50

The European Union’s AI Act, Regulation 2024/1689, contains a transparency article that matters directly to advertisers. Article 50 requires providers of generative systems to mark synthetic audio, images, video and text in a machine-readable way. It also requires anyone deploying a system to create a deepfake to disclose that the content is artificial.

For clearly artistic, satirical or fictional work, the deepfake duty shrinks to disclosing that such content exists, in a way that does not spoil the piece. These obligations apply from August 2, 2026. You will increasingly see small labels or badges on ads aimed at European audiences, along with metadata that most viewers never notice.

Other rules and the problem of AI washing

Outside Europe, regulators mostly apply existing advertising law. In the United States, the Securities and Exchange Commission settled charges in March 2024 against two investment advisers for exaggerated claims about their use of AI, a practice it called AI washing. Consumer protection agencies have also warned companies against promising AI features that do not exist.

Canada has no dedicated statute yet. A proposed Artificial Intelligence and Data Act was part of Bill C-27, and it never became law before Parliament was prorogued in January 2025. Canadian advertisers therefore follow the Competition Act and voluntary industry codes on misleading claims. The gap is worth watching, since Canadian brands running campaigns in Europe already face Article 50.

Examples of what still works, and what does not

Looking across these examples, a pattern holds. AI spots that admit they are AI, or that play with the strangeness, fare better than ones that pretend to be warm human stories. Nostalgia is the riskiest ground of all. Remaking a beloved holiday ad with software invites comparison with the original, and most viewers prefer the original.

Craft has not disappeared from the conversation either, whatever the tools promise. Typography, colour and layout still decide whether an ad reads well, which is why the graphic design trends for 2026 lean toward texture and visible handwork. Whether Coca-Cola’s trucks will roll out of a render farm again next November is an open question.