Although AI’s environmental impact cannot be disregarded, some use cases of the technology may actually help address the climate crisis. Cities are now looking to add AI into operations, but they have already been working to reduce landfill waste through technology.
Last month, the city of Centerville, Ohio — a suburb of Dayton — launched an AI-powered program aiming to address and reduce recycling contamination. The pilot is using camera-based technology to identify contamination amongst the contents of recycling carts, in real time. Items that do not meet local recycling guidelines are flagged. Then, the city can contact residents with a personalized notecard offering guidance to mitigate future errors.
The city will evaluate the AI system’s impact, leveraging data to inform future program improvements.
Centerville is not alone in using AI technology this way; San Francisco announced the deployment of three installations of Oscar Sort, a zero-waste AI recycling assistant, at the city’s Ferry Building in May to support its zero-waste goals. This builds on the tool's deployment at the Exploratorium.
“[T]he Ferry Building sees thousands of people from all around the Bay Area and beyond passing through its doors each day — many of whom are unfamiliar with SF's three-bin system,” Joseph Piasecki, the San Francisco Environment Department’s public affairs and policy coordinator, said via email.
The technology identifies the materials users are going to discard and informs them of the correct bin to use — compost, recycling, or landfill. If users make a mistake, Oscar informs them; some users have even reached in and re-sorted an item upon Oscar’s prompting, Piasecki said.
Intuitive AI, the creator of Oscar Sort, shared the video below of the technology in action, demonstrating how users can scan their item to receive disposal instructions.
The system even engages users by gamifying the recycling experience, giving them the chance to play trivia and even offering rewards for proper waste disposal.
The Oscar Sort system is instructed and programmed to recognize items common to San Francisco and categorize them under one of these three labels, Piasecki said. The programming is stored on a thumb drive and the machine’s camera is aimed low to protect people’s privacy from facial scanning, which sometimes can pose risks.
Notably, while generative AI requires significant amounts of energy, Piasecki emphasized that there is no immediate generative AI component happening at these stations, therefore requiring “far less energy.”
The ultimate goal is to improve sorting and minimize the contamination of recycled and composted materials. The technology’s impact in the city will be made evident based upon public locations’ refuse audits and feedback from the city’s refuse hauler, Recology, Piasecki said.
The implementation’s focus is on educating residents and using innovation to reduce landfill waste. In this case, he said, assistive AI contributes to making the experience of waste disposal “interactive, fun, and relevant” for San Franciscans.
“We would encourage other cities to explore any resources that would help achieve their zero-waste goals,” Piasecki said.
