AI, Games, and the Environment:
A Responsible User’s Guide

Principles in Practice

These principles in practice are intended to help turn the guiding principles into action and have been developed according to the three thematic groups of the principles.

‍

Engagement

Contributing to a constructive dialogue with model developers and providers to address the transparency gap​.

Inconsistent reporting and limited transparency add to the challenge of tracking outcomes and impacts of AI use, including environmental impacts. Studios can play a role by communicating expectations, requesting relevant information, and supporting industry-wide improvements.

  • Engage constructively with peers in the games industry on sustainability, transparency, and reporting.
  • Request information that supports informed procurement and governance decisions.
  • Support cross-industry efforts to develop common measurement and reporting frameworks.
  • AI service providers data request checklist
  • Lifecycle energy and carbon intensity of the models being offered, i.e., Wh per token, Wh per image, Wh per video, Wh per task, etc. and gCO2e per token, gCO2e per image, gCO2e per video, gCO2e per task, etc.
  • Total energy consumption and carbon emissions related to your use of the AI service on a monthly or annual basis, using GHG Protocol or equivalent carbon accounting standards.
  • Mix of electricity generation sources in the supply chain of the AI service, including exposure to behind-the-meter gas turbine powered data centres.
  • Transparency around water consumption and water management in the supply chain of the AI service.

Training: Stay informed, curious and agile and teach others what you have learned​

‍Stay informed, curious and agile and teach others what you have learned​

The development of AI and its capabilities are evolving rapidly. Transparency, resulting in uncertainties of environmental impacts, are also part of it. Studios should foster continuous learning, knowledge sharing, and also fostering good practice.

  • Encourage your team to share lessons learned from AI projects and experiments, while maintaining appropriate safeguards: case studies on real-world usage and learnings can be incredibly valuable and insightful for others.
  • Create opportunities and internal mechanisms for cross-functional or cross-divisional discussions on AI use and its impacts.
  • Provide ongoing education or internal training on AI tools, including its use-specific capabilities and limitations.
  • Monitor developments in regulation, standards, best practices, and lessons learned.
  • Update internal guidance along the process.

‍