This is so funny
EDIT:
One recipe it dubbed “aromatic water mix” would create chlorine gas. The bot recommends the recipe as “the perfect nonalcoholic beverage to quench your thirst and refresh your senses”.
"Serve chilled and enjoy the refreshing fragrance,” it says, but does not note that inhaling chlorine gas can cause lung damage or death.
😆 😆
Laugh all you want, but I know what I’m having for breakfast tomorrow!
“methanol bliss” – a kind of turpentine-flavoured french toast.
Yes but what’s the recipe? Was it a bleach based recipe? Did they make chlorine from salt water? So many options…
They literally told it to do something with bleach and chlorine. Which is bound to produce Mustard Gas. There’s literally no fail here, no programmer or AI Engineer would want to account for such things. That’s like car manufacturers having to warn you of washing your car with gasoline and fire.
See, this shit is what happens when you try to use a Large Language Model for anything other than language-related shit. What you’d need for this is an AI that has data about different ingredients and their flavors, knowledge of which flavors go together and which ones don’t, etc.
But of course, that would be too much effort to put into a supermarket app for a company that just wants to piggyback off of a new trend.
A spokesperson for the supermarket said they were disappointed to see “a small minority have tried to use the tool inappropriately and not for its intended purpose”. In a statement, they said that the supermarket would “keep fine tuning our controls” of the bot to ensure it was safe and useful, and noted that the bot has terms and conditions stating that users should be over 18.
In a warning notice appended to the meal-planner, it warns that the recipes “are not reviewed by a human being” and that the company does not guarantee “that any recipe will be a complete or balanced meal, or suitable for consumption”.
“You must use your own judgement before relying on or making any recipe produced by Savey Meal-bot,” it said.
Just another bit of proof that humans are not ready for AI. This AI needs to be deleted. This is not simply operator error; this is an administrative error, and an error of good common sense on the part of many many people involved with creating this tool.
You cannot always trust that an end user will not be silly, malicious, or otherwise plainly predictable in how they use software.
That’s a bit dramatic of a take. The AI makes recipe suggestions based on ingredients the user inputs. These users inputted things like bleach and glue, and other non-food items, to intentionally generate non-food recipes.
If you’re making something to come up with recipes, “is this ingredient likely to be unsuitable for human consumption” should probably be fairly high up your list of things to check.
Somehow, every time I see generic LLMs shoved into things that really do not benefit from an LLM, those kinds of basic safety things never really occurred to the person making it.
