Microsoft AI Data Centers in Mexico Cause Power Outages and Water Shortages

The article explicitly links the operation of AI data centers to significant environmental and social harms, including depletion of water resources, increased electricity demand leading to outages, and consequent health and economic impacts on local communities. These harms fall under harm to communities and the environment. The AI systems depend on these data centers, so the harms are indirectly caused by AI system use. The article also describes protests and regulatory responses, but the primary focus is on the harms already occurring. Thus, the event meets the criteria for an AI Incident rather than a hazard or complementary information.

The article explicitly connects the operation of AI data centers to real, material harms experienced by local populations, including health problems and disruptions to essential services like water and electricity. The AI systems are central to the data centers' purpose, and their resource demands have contributed to these harms. Although the harms are indirect, the causal link is clear and supported by multiple examples and testimonies. Therefore, this event meets the criteria for an AI Incident rather than a hazard or complementary information.

The article explicitly links the growth of AI data centers to increased energy and water consumption, which has caused or worsened power outages and water shortages affecting local populations' health and livelihoods. These data centers are integral to AI system operation, so their environmental and social impacts are consequences of AI system use. The harms described include health issues, disruption of essential services, and harm to communities, fitting the AI Incident definition. While the article notes difficulties in directly attributing all harms to specific centers, the overall pattern and expert opinions support a causal link. The event is not merely a potential risk (hazard) or a governance or research update (complementary information), but a documented case of harm associated with AI system infrastructure.

Microsoft data centers host AI services and require massive energy and water resources to operate. The article links these centers to real harms in local communities, including health impacts and infrastructure disruption. Despite Microsoft's denial, the credible reports of outages and water scarcity caused or worsened by the centers' resource consumption meet the criteria for harm caused directly or indirectly by AI systems' use. The harms are materialized, not just potential, and affect health, communities, and critical infrastructure, fitting the AI Incident definition.

The Microsoft data center supports AI development and operation, thus involving an AI system. The event reports actual harms to people and communities (power outages, water scarcity, illnesses) directly linked to the data center's opening and operation. These harms fall under categories (a) injury or harm to health and (d) harm to communities. Hence, the event meets the criteria for an AI Incident as the AI system's use has indirectly led to significant harm.