Sam Altman defends AI resource usage: Water concerns 'fake,' and 'humans use energy too'

During the recent India AI Impact summit, OpenAI's CEO Sam Altman took a stand against criticisms surrounding the environmental impact of artificial intelligence, particularly focusing on the water consumption of data centers. He labeled the assertions that AI systems like ChatGPT consume excessive water per query as "completely untrue" and insisted that such claims bear "no connection to reality." Altman explained the role of water in data centers, which traditionally use it to cool hardware and prevent overheating. While advancements in technology have led to some facilities reducing water usage or eliminating it altogether, a report from Xylem and Global Water Intelligence raised concerns that water use for cooling could triple in the next 25 years due to the increasing demand for computing. While dismissing the water usage concerns, Altman acknowledged the significance of energy consumption in the AI debate. He emphasized the necessity for a transition to more sustainable energy sources like nuclear, wind, and solar to meet the growing demands of AI technology. Altman noted that while the energy consumed overall is a valid concern, the efficiency of AI in terms of energy use per query is already comparable to that of humans. In response to comments from Microsoft co-founder Bill Gates regarding the efficiency of the human brain, Altman argued that the energy required to train an AI model should be fairly compared to the extensive resources used to develop human intelligence over two decades. He pointed out that once an AI model is trained, the energy needed to generate responses—known as inference—is significantly less than the training phase itself. Altman’s remarks have ignited discussion on social media, especially regarding the implications of AI potentially replacing human roles. Critics, including Zoho Corporation’s co-founder Sridhar Vembu, expressed discomfort with equating AI capabilities to human intelligence. The conversation around AI's resource demands is becoming increasingly relevant as companies and governments invest heavily in new data centers to support AI operations. A recent report indicated that electricity consumption by data centers in 2023 has reached levels comparable to that of entire nations like Germany or France. In light of these developments, some local governments are accelerating efforts to enhance energy supply, despite environmentalists warning that such initiatives could conflict with global sustainability goals. Communities, particularly in the U.S., have pushed back against new data center projects over concerns about their impact on electricity costs and grid stability. For instance, the San Marcos City Council recently rejected a significant $1.5 billion data center proposal after a prolonged public opposition campaign. Amid these challenges, tech leaders, including Altman, advocate for a shift towards diverse and renewable energy sources for future data centers.