TLDR: MIT students Ayah Mahmoud and C Jacob Payne have developed Kitchen Cosmo, an innovative AI-powered device that scans ingredients and generates personalized recipes, which are then printed. The device utilizes GPT-4o and features a tactile, screenless interface, aiming to transform cooking into a co-creative experience.
Cambridge, MA – A groundbreaking culinary innovation, ‘Kitchen Cosmo,’ has been unveiled by MIT students Ayah Mahmoud and C Jacob Payne. Developed within the ‘Interaction Intelligence’ course, taught by Professor Marcelo Coelho, this speculative AI cooking device challenges conventional smart kitchen technology by fostering a co-creative relationship between users and artificial intelligence.
Kitchen Cosmo operates by scanning available ingredients using an integrated webcam. It then leverages GPT-4o, a multimodal large language model capable of processing both images and text in real-time, to generate personalized recipes. A single API call translates these diverse inputs into a context-specific recipe, which is then printed out via an embedded thermal printer.
One of the device’s most distinctive features is its entirely screenless and tactile interface. Eschewing modern digital displays and voice assistants, Kitchen Cosmo relies on a series of knobs, sliders, and physical rituals for user interaction. This retrofuturistic design, with its striking red shell and cylindrical recipe archive, draws inspiration from the 1969 Honeywell Kitchen Computer. However, unlike its non-functional predecessor, Kitchen Cosmo is fully operational and designed for improvisational, rather than prescriptive, cooking.
A key element of its human-centered approach is the ‘Cooking Mood’ dial. This unique input allows users to infuse emotional character into their recipes, with settings such as ‘Nostalgic,’ ‘Spectacle,’ and ‘Surreal.’ This innovative feature influences the structure, tone, and logic of the generated recipes, moving AI beyond mere standardization towards interpretation and playful creativity.
Kitchen Cosmo is also part of a broader research initiative into ‘Large Language Objects’ (LLOs), which are physical artifacts that integrate generative models into tangible, interactive systems. While early user testing identified limitations related to cultural bias and recipe reliability, the project’s framework is designed for future expansion, including user-specific training, cookbook-informed datasets, and deeper integration of diverse cultural food knowledge.
Also Read:
- San Jose Pioneers AI Grant Program to Tackle Maternal Health, Food Waste, and Satellite Data
- AI’s Dual Impact: Job Market Transformation, Security Concerns, and Future Prosperity Debated by MIT Technology Review
Ayah Mahmoud and C Jacob Payne envision Kitchen Cosmo as a visible, embodied collaborator in the everyday ritual of cooking, offering a critical vision for how artificial intelligence can support domestic life.
