Google’s AI is breaking recipe sites

Why food blogs are losing traffic?

Google’s AI is now cooking dinner for people.

Since March, Google’s AI Mode has been generating recipes directly in search results, often by stitching together instructions from 8 to 10 different creators into one simplified version. Recipe writers report steep traffic drops, including one site that lost 80% of visits in two years, right as the holiday season should be paying the bills.

I think this lands harder than past search changes because it breaks trust, not just traffic. When a recipe turns into a thin list of steps stripped of judgment, testing, and context, readers stop learning and creators lose the reason to show up.

Some bloggers are shifting to video or subscriptions, but most cannot afford to disappear from Google entirely. Cookbooks are selling better this year, yet even those are being scraped and resold by bots.

At some point, standing in a kitchen at 6pm, people will notice when the food keeps coming out wrong.

We bought the tools, skipped the people

Most companies are spending on AI like it installs itself.

Deloitte CTO Bill Briggs says companies are putting 93% of their AI budgets into technology and only 7% into people, training, and workflow redesign. He shared the stat while discussing Deloitte’s latest Tech Trends report from New York, as firms move from small pilots toward AI at scale.

This rings true because it matches how enterprises always behave under pressure. Buying software feels decisive, while changing jobs, incentives, and habits feels slow and risky, so leaders default to what fits neatly into a budget line.

The cost shows up as falling AI usage, down 15%, and rising “shadow AI,” with 43% of workers bypassing approved tools because the unofficial ones are easier. Trust collapses, even as spending keeps rising.

If the tools are ready but the people are not, the real question is how long companies will keep paying for systems their employees quietly avoid using.

Use the tool, accept the risk

The lawyers are no longer saying no.

On the Digiday Podcast, IP lawyer Rob Driscoll explains how brands have moved from banning generative AI to requiring it, even as copyright and trademark risks stay unresolved. Disney’s deal with OpenAI, which allows AI-generated Mickey Mouse and Darth Vader content under tight controls, shows how much legal effort it takes to make this workable.

What stands out is how normalized uncertainty has become. Brands know AI output can accidentally copy existing work or weaken trademark protection, and many are proceeding anyway because speed and scale matter more than clean ownership.

The safer players are paying for private, enterprise AI tools so their inputs do not train shared models. Others are accepting that they may not fully own what they publish, as long as it performs and does not trigger a lawsuit.

If using AI is now mandatory, the real decision is how much legal risk a brand is willing to live with when the lights are on and the campaign is live.

Prohuman team

Free Guides

Explore our free guides and products to get into AI and master it.

All of them are free to access and would stay free for you.

Feeling generous?

You know someone who loves breakthroughs as much as you do.

Share The Prohuman it’s how smart people stay one update ahead.