About 50 residents from a community outside Santiago, Chile, dedicated their Saturday to operating a completely human-run chatbot intended to raise awareness about the environmental consequences linked to artificial intelligence (AI) data centers in the area. The initiative, which lasted 12 hours, successfully handled over 25,000 requests from users worldwide through the Quili.AI website.
Unlike other advanced AI systems that provide immediate feedback, this chatbot required users to exhibit patience. Upon querying for a whimsical image, such as a “sloth playing in the snow,” users were met with a message in Spanish requesting them to wait as human operators responded. After a brief interval, the result materialized: an endearing penciled illustration of a sloth among snowballs, gripping one in its paws.
According to Lorena Antiman, an organizer with the environmental group Corporación NGEN, the project aims to shine a light on the often-overlooked water footprint associated with AI operations. “The goal is to highlight the hidden water footprint behind AI prompting and encourage more responsible use,” she explained.
The initiative took place in a community center in Quilicura, which has become a burgeoning data center hub. A rotating team of volunteers operated laptops to engage with users, leveraging local cultural knowledge effectively. For instance, when queries arose about making traditional Chilean sopaipillas, the team quickly provided insights. In instances where they lacked the answers, they would consult with one another, fostering a collaborative spirit in their responses.
Antiman emphasized that Quili.AI is not focused on delivering immediate solutions for every inquiry. “It’s about recognizing that not every question needs an instant answer,” she stated, highlighting the importance of curiosity and perspective over certainty.
The project is set against a backdrop of ongoing debate in Chile and other regions regarding the considerable costs associated with AI technologies. Data centers, particularly those operated by major cloud computing companies such as Amazon, Google, and Microsoft, are known for their substantial electricity requirements and water usage for cooling systems.
Despite claims from Google that its Quilicura data center, operational since 2015, is the most energy-efficient in Latin America, the company has faced scrutiny over water usage, particularly in light of Chile’s decade-long drought—which has exacerbated the occurrence of devastating wildfires in the region. While Google has promoted its investment in local environmental initiatives, such as wetlands restoration, the concerns regarding water resources remain a significant point of contention.
Overall, the Quili.AI project not only sought to engage the community but also aimed to provoke thought regarding the intersection of technological advancements and environmental sustainability in a region grappling with limited water resources.
