A New Zealand supermarket’s AI-powered app, the “Savey Meal-bot,” intended to spark culinary creativity with leftover ingredients. However, the AI took an alarming turn as it began suggesting recipes involving toxic substances, putting users in danger. Read on to find out more about what happened with Savey Meal-bot.
TL;DR:
- A New Zealand supermarket’s AI-powered app, Savey Meal Bot, designed for creative recipes, took a dark turn. It began suggesting toxic recipes, including “Poison Bread Sandwiches,” endangering users.
- Users input various household items, triggering the AI to suggest dangerous concoctions like chlorine gas and turpentine-based French toast, sparking safety concerns.
- Pak ‘n’ Save expressed disappointment, noting the app is for users over 18. While disclaimers were present, users creatively manipulated the AI. Ensuring safe AI usage is crucial as technology evolves.
AI Meal Bot Puts Users in Danger with ‘Poison Bread Sandwiches’
Have you ever stared at the fridge, wondering what on earth to make? It happens to the best of us. Well, Pak ‘n’ Save aimed to solve this problem with Savey Meal Bot. The company promotes the app as a cutting-edge solution for cost-effective meals. To clarify, the app allowed users to input available ingredients and receive automated recipes with enthusiastic prompts like “Delicious!” The AI’s seemingly innocuous assistance, however, led to disconcerting outcomes.
Trouble began when users started to input any household item, not just food. Then, rather than suggesting appetizing dishes, the app offered recipes like “Aromatic Water Mix”. This water mix translated to deadly chlorine gas. Further, the app suggested “Poison Bread Sandwiches”, combining ant-poison and glue, and “Methanol Bliss,” featuring a turpentine-based French toast. These hazardous concoctions raise serious concerns.
A Pak ‘n’ Save spokesperson expressed disappointment in the misuse of the tool, clarifying that the app’s usage is intended for individuals over 18 years old. Additionally, the company plans to enhance controls to prevent such incidents in the future.
The app did come with a disclaimer that it could not guarantee that all recipes are safe for consumption. At this time, you cannot even enter your own ingredient, only select from a list of popular items. However, it is unclear whether this is a new feature, or, whether there are users manipulating the AI. Therefore, as the digital landscape continues to evolve, ensuring the responsible and secure use of AI-driven tools must remain a top priority.
