Everyone is in danger! And again, artificial intelligence surprises: unverified data, a coincidence of circumstances, and human curiosity - what it can lead to.
According to The Guardian in New Zealand, a supermarket application with artificial intelligence suggested meal plans for customers and helped use leftover products. But some have noticed that the dishes are a little unusual, such as recipes for deadly chlorine gas and "poison bread sandwiches" and fried potatoes to repel mosquitoes.

As customers added items from their shopping list to the app, even less attractive recommendations were made, including "flavored water mix" (chlorine gas).

I will have to ask my wife before dinner where she got the recipe for today.

But everything is not as bad as it seems at first glance.
After social media outrage over the recipes, the supermarket said it would continue to fine-tune the bot's controls to make sure it's safe and useful
It's worth noting that the planner now includes a warning that the recipes it generates have not been tested by humans and may not be fully balanced or suitable for consumption.

There is a consensus among EU countries and the USA regarding the regulation of AI. Additionally, in my view, alongside warnings about the use of unverified data with certain AI models, caution about slow usage should also be included.