Artificial Intelligence faces scrutiny over hallucinations and environmental impact

Generative artificial intelligence is facing significant challenges regarding technical accuracy and environmental sustainability. Research from Stanford RegLab indicates that advanced legal AI search tools exhibit hallucinations in 17% to 33% of queries. A critical issue identified is “erroneous foundation,” where AI cites real legal precedents but attributes incorrect conclusions or rulings to them.

In response to these risks, the European Union is implementing the AI Act, which mandates transparency requirements such as labeling synthetic content and deepfakes.

Beyond accuracy, the physical infrastructure required for AI is drawing regulatory attention. Data centers require massive amounts of electricity and water for cooling. In Spain, the government has approved regulations for large data centers to ensure maximum water and energy efficiency, requiring electricity consumption to be linked to new renewable energy sources. While these regulations aim for sustainability, some investors express concerns regarding project competitiveness and viability.