The AI Learning Hub
Published 1 November 2026

AI and Marketing: What Changed, What Didn't, and the Rules That Apply

Marketing was one of the first jobs AI changed visibly. Within a couple of years, tools that write copy, generate images, edit video and analyse campaigns went from novelties to standard equipment. If you work in marketing, you have almost certainly used them. If you are a customer, you have almost certainly seen their output, whether you noticed or not.

This post looks at where AI genuinely helps marketers, where it backfires, how it is changing the way customers find brands, and the rules that already apply. It is dated 1 November 2026.

Where AI genuinely helps

First drafts, in volume. Product descriptions, ad variations, email subject lines, social posts. AI is very good at producing many reasonable versions quickly, which suits a job that has always depended on testing variations.

Images and video without a shoot. Mock-ups, background scenes, product shots in different settings and quick video edits that once took a studio day can now take minutes. Small businesses benefit most, because they could never afford the studio.

Localisation. Adapting a campaign into other languages and markets used to be slow and expensive. AI translation and adaptation make it cheap. It still needs a native speaker to check tone and cultural fit.

Making sense of data. Summarising thousands of customer reviews, survey answers or support tickets into themes is now fast. This may be the most underrated use: it tells you what customers actually think, in their own words.

Personalisation. Tailoring a message to a customer’s situation used to mean a few broad segments. AI makes it practical to vary content far more finely.

Where it backfires

Everything starts to sound the same. When every competitor uses the same tools with similar prompts, the output converges: the same upbeat tone, the same structure, the same phrases. Audiences learn to recognise it and scroll past. Generic content is cheaper than ever, which is exactly why it is worth less.

Confident mistakes. AI will happily invent a product feature, a statistic or a customer quote. In marketing, a made-up claim is not just embarrassing. It can be false advertising. Every factual claim needs checking against the real product.

Brand damage. Several brands have faced public backlash for AI-generated images with obvious errors, or for replacing human creatives in campaigns about authenticity. Audiences are forgiving of AI used well and harsh on AI used carelessly.

Personalisation that feels like surveillance. There is a line between “relevant” and “how do they know that?”. Crossing it costs trust, and in many places it can break data-protection law.

Customers now find you through AI too

Marketing has always followed where customers look, and customers increasingly ask AI assistants. “Which project management tool suits a team of five?” might now go to a chatbot, not a search engine, and the answer names a handful of products.

That creates a new question for marketers: what does AI say about your brand? AI assistants form their view from what is written about you across the web: your own site, reviews, comparisons, forums and press. You can’t buy your way into that answer the way you can buy an ad. You earn it by being clearly described, well reviewed and genuinely recommended. Our companion post, AI and Search, covers what this means for being found online.

The rules that already apply

There is no special “AI marketing law” in most places, but existing rules apply in full, and some new ones matter.

  • Fake reviews are banned, however they are made. In the United States, the Federal Trade Commission’s rule on consumer reviews and testimonials, finalised in August 2024, prohibits fake reviews and testimonials, explicitly including AI-generated ones, and allows civil penalties against companies that knowingly break it. Labelling a fake review as AI-written does not make it legal. The problem is that it’s fake.
  • False advertising is false advertising. If AI writes a claim and you publish it, it is your claim. You need evidence for it, just as before.
  • Endorsements must be disclosed. Paid, gifted or employee endorsements need clear disclosure. That applies to AI-generated “influencers” and spokespeople as well as real ones.
  • Transparency rules are arriving. The EU AI Act includes transparency duties around AI-generated content, such as making deepfakes and synthetic media identifiable, and major ad and social platforms already ask for AI-made content to be labelled in some cases. Check the current requirements in each market you advertise in. Our post on the EU AI Act’s moving deadlines explains why the dates keep shifting.
  • Personal data rules still apply. Feeding customer data into AI tools for personalisation is still processing personal data. Privacy law and your own privacy policy apply.

Using AI in marketing well

  1. Use AI for volume, humans for voice. Let AI draft and vary. Let a person decide what the brand actually sounds like and what is worth saying.
  2. Check every factual claim against the real product, the real data or the real customer quote before it goes out.
  3. Start from real insight. AI-generated content built on a genuine customer insight beats any amount of generic output. Use AI to find that insight in your own data.
  4. Be open about meaningful AI use. You don’t need to label every AI-assisted subject line, but synthetic people, voices and images that could be mistaken for real deserve disclosure.
  5. Keep customer data out of tools you haven’t approved. Use business tools with clear data terms, not personal chatbot accounts.

Where to go deeper

AI Essentials explains what these tools are good and bad at, and why they make things up. Prompting will make your first drafts noticeably better, and AI for Legal covers the legal side of AI in more depth. All free.

Related free courses