Image by Anton Nazaretian, from Unsplash

AI Tools Help Restaurants Slash Food Waste And Costs

• Written by Kiara Fabbri Former Tech News Writer
• Fact-Checked by Sarah Frazier Former Content Manager

AI may be changing how restaurants address food waste problems.

In a rush? Here are the quick facts:

• Food waste produces 8% of global carbon emissions.
• AI tool Winnow identifies 1,000+ foods as they’re discarded.
• Four Seasons Costa Rica cut food waste by 50% in 8 months.

The restaurant and hospitality sector loses more than $100 billion a year in food waste, as reported by Business Insider (BI). According to ReFED’s 2025 report , U.S. businesses alone had surplus food worth $108 billion in 2023. Beyond the financial hit, food waste accounts for 8% of global carbon emissions, as noted by BI.

In order to overcome this issue, BI reports that many businesses are resorting to AI solutions. For example, Winnow is an AI tool that helps restaurants monitor their waste. It does this using scales and motion-sensing cameras, which identify more than 1,000 discarded food items, providing chefs with immediate data and weekly performance reports.

BI reports Winnow use in more than 90 countries, assisting restaurants in portion control, leftover utilization, and meal planning.

For example, Four Seasons Peninsula Papagayo in Costa Rica cut food waste by 50% in eight months, as a result of Winnow. “It’s simply part of the basic daily operation,” said executive chef Emiliano Rabia Sottil, who now donates extra meals to food banks and reuses trimmings for stocks.

In another case, Hotels in Ireland have also reached significant waste reduction results. For example, Ashford Castle reached 69% food waste reduction in 2024, while The Lodge achieved a 90% decrease.

“It’s brought about a lot more awareness on our waste,” said Liam Finnegan, executive head chef at Ashford Castle, as noted by BI. Chefs there now turn croissants into bread pudding, carrot tops into pesto, and chicken skin into crackers.

Staff are embracing the change, too. “It’s very tangible and engaging for young people,” said Jonathan Keane, executive head chef at The Lodge, as reported by BI.

For many chefs, technology is more than a tool. “The idea is to create that culture,” Rabia Sottil said to BI, pointing to a future where cutting waste is as important as creating delicious food.

Beyond hotel kitchens, similar efforts are underway in the UK . Indeed, the UK food-tech company Zest is working with Nestlé and other partners to test an AI system that identifies surplus food products to find alternative distribution channels.

The two-week trial of this tool resulted in an 87% reduction of food waste, potentially preventing the loss of 1.5 million meals nationwide.

Recent research supports the effectiveness of AI-based waste-tracking systems in reducing food waste across hotels and restaurants. A new study found that implementing AI devices reduced food waste by 23–51% per meal, while organizations saved between 23–39% of their costs.

The intervention program raised employee understanding, which led to various corrective actions, like adjusting portion sizes and preparing buffet items only when needed.

With these results in mind, universities , businesses, and policymakers need to establish responsible implementation methods for these tools to achieve lasting social and environmental benefits.

Image by Mufid Majnun, from Unsplash

AI Virtual Patient Trains Future Healthcare Workers

• Written by Kiara Fabbri Former Tech News Writer
• Fact-Checked by Sarah Frazier Former Content Manager

Medical students at Coventry University will now practice through a virtual patient system powered by AI, aimed at developing their skills through realistic patient simulations.

In a rush? Here are the quick facts:

• Coventry University developed an AI avatar named Sarah for medical training.
• Sarah provides natural, personalized answers, unlike scripted virtual patients.
• AI reduces creation time from 100 hours to about 30 minutes.

The system uses advanced large language models which generate real-time conversations with an AI avatar, named Sarah. This tool was developed to help future healthcare workers build confidence before stepping into real-world practice.

Unlike earlier versions that relied on scripted responses, this new system uses advanced large language models guided by ChatGPT-style prompts to give personalised, natural answers.

According to the PhD student who led the project, Juliana Samson, the AI system produces more authentic interactions between users and the avatar.

“Sarah’s answers are more consistent with the intended history that we want the students to get out of the scenario,” she explained to the BBC . Samson, who previously worked as a physiotherapist for twenty years, emphasized that the tool provides a “safe space” for students to practise without the pressure of treating a live patient.

Students also praised the system. “Students have told us it will help to build their confidence, especially for those who are maybe earlier on in the programme, and they enjoyed the activity,” Samson said to the BBC.

Indeed, experts believe this approach reflects a wider trend in education. Research on AI in virtual reality (VR) environments shows that personalized AI narration can significantly boost engagement.

Research shows that tailored AI interaction boosts student participation. The research on VR learning demonstrates that tailored AI interaction leads to a 64% rise in user participation while keeping cognitive load manageable. The research indicates that adaptive AI systems such as Sarah would both enhance student preparation and sustain their motivation over time.

The technology has another big advantage: speed. What once took 100 hours to design can now be drafted in just 30 minutes. “With this artificial intelligence we have been able to create an avatar that actually behaves like a patient and can be created quickly,” Samson said in a University press release .

The AI allows students to ask questions, take patient histories, and receive feedback points, all without risk of upsetting a real person.

The project is being developed with AI firm PCS. “We are thrilled to provide Coventry University our most intelligent Spark virtual patient AI, now driven by an in-house trained generative large language model for never-before-seen realism in communication training,” said PCS CEO Balazs Moldovanyi, as noted in the University press release.

Samson hopes to expand its use into occupational therapy, physiotherapy, and other healthcare fields.