Why SAP screens every AI project for ethical and environmental risks

This is corporate software news, not something that changes your home energy options directly, but it's a window into how big tech companies handle the environmental cost of artificial intelligence.

SAP, a major business software maker, has an ethics policy that requires every new AI project to be screened for environmental impact before it's built, alongside privacy and human rights concerns. Projects get a risk score, and anything flagged as high risk goes to a special committee for review. Few major tech companies do this: Amazon, Google and Microsoft have AI ethics rules but none include criteria for greenhouse gas emissions, according to the company. Salesforce and IBM are among the few others that do.

SAP says it only adds AI to a product when it delivers real value, to avoid wasting computing power. The company has also worked to use smaller, more efficient AI models, automatically route tasks to whichever model uses the least energy, and match the electricity used at its own data centers with renewable power. It says these steps cut average emissions per AI query by 60 percent between 2024 and early 2025.

None of this changes what's available for home energy upgrades or rebates. But it matters for anyone tracking how the tech industry's AI boom affects overall electricity demand and emissions, since data centers increasingly compete with homes for grid capacity and clean power.