Home: Home > Features > Content


Principles on AI for Climate Action


Principles on AI for Climate Action

Addressing global climate change is essential for the sustainable future of humanity and ecology. The stakeholders of Artificial Intelligence (AI) Science and Technology should deeply collaborate with stakeholders of climate change and ecosystems to contribute to the implementation of the global climate agenda, such as the United Nations Framework Convention on Climate Change (UNFCCC), the Paris Agreement, and the UNESCO Strategy for Action on Climate Change. With international ethical principles such as those enshrined in the UNESCO Recommendation on the Ethics of Artificial Intelligence, the UNESCO Declaration of Ethical Principles in relation to Climate Change, etc. as the foundation, the proposed principles here serve as cross domain efforts on development and ethical principles of AI for climate actions. Aimed at all stakeholders in AI and climate action, including management, Research and Development, use, deployment, investment, etc., these principles aim to provide relevant parties with principles of AI development and governance to address climate change through deep synergy between technology, ethics and governance, and contribute to the steady advancement of the global climate agenda.

Values and Principles

For human and ecology good

AI technology and its applications should not only serve for the development of human society, but should contribute to the symbiosis of humankind, ecological systems and the environment. Hence they should also generally be beneficial to the control of climate change and the improvement of the ecological environment, and should be used as an enabling technology to support the realization of the overall goals of global climate action and the climate agenda, and contribute to the realization of carbon peaking and carbon neutrality goals.

Energy conservation

The development and use of AI should reduce its own energy consumption as much as possible while meeting specific needs. AI applications with high energy consumption should follow the necessary principles and alternative solutions with relatively low energy consumption should be actively explored. For example, simplifying the AI model within the acceptable range of loss, optimizing the model training method, and co-design of software and hardware to achieve the expected level of intelligence with low energy consumption. The potential approach includes but is not limited to smart and low energy consumption. Considering the type and application characteristics of electric power comprehensively, and planning for rational use, where appropriate, green power are encouraged to be used, and efforts on reducing the cost of power storage should be continuously made. It is encouraged to train AI models and algorithms and deploy AI applications in areas where electric power is relatively cheap, and adopt solutions that consume less energy for server cooling as much as possible.

Privacy protection

In promoting the application and activities of AI for climate action, it is necessary to ensure people's rights to privacy and informed consent. Personal privacy related data cannot be illegally obtained in the name of climate control. Where data is used for purposes of climate action, data subjects should retain choice and control in how the data is managed.

Fairness and justice

AI should be transparent and objective when it contributes to climate action. When it is used to assess, analyze and predict the impact of countries, regions and industries on climate change, their characteristics and development stages should be considered to avoid introducing bias. AI should contribute to the attention and evaluation of potential additional damages suffered by vulnerable groups in climate change, as well as the assessment of the negative impact of new technology revolutions on climate change, so as to avoid exacerbating inequality between countries, regions and social groups due to climate change. AI technologies and systems related to climate change control are encouraged to be opensource and shared. We should actively empower low- and middle-income countries, and regions with lower development status on AI to accelerate the realization of global climate goals.

Promote education, training and employment

AI should contribute to the education and training on climate change and climate action, to promote government, industry and public awareness, to promote the technical implementation of climate action policies, and to the guidance for people whose employment is affected by climate change and climate action. AI should contribute to incubate green industries and to promote reemployment.

Sharing and cooperation

Climate change is a global issue, and human society should collaborate as a whole to handle it, supported by technologies like AI. Climate action requires the establishment of an international cooperation platform empowered by data and AI to promote global collaboration. The formation and use of a global climate change data platform, a global greenhouse gas emission monitoring platform, a global real-time carbon trading data platform, and a global real-time energy trading data platform should be promoted. Governments, academia, industry, users, professionals in climate change and AI, and other related fields should deeply collaborate to ensure that AI contributes positively to climate action and the achievement of global goals.

Recommendations for Action

The following recommendations aim to promote AI-powered climate action and provide reference to AI researchers, practitioners, and policymakers. The actions recommended below should comply with national, regional, and international ethical principles and specifications on AI, environment, and the principles listed in this document above. The list of recommendations will continue to be enriched as the status of climate change and climate action evolve.

Facilitate climate analysis and forecasting

AI can be used to assist in monitoring the causes and states of climate change, and through computational modeling and simulation, it can contribute to the understanding of climate mechanisms and forecast climate change trends, assist avoiding climate risks and crisis, such efforts should especially provided to low- and middle-income countries for early warning of climate disasters. It can be used to track the sources and impacts of greenhouse gases, to analyze and predict the impact of climate change and climate policies on the economy, politics and people's livelihood.

Promote energy conservation

AI can be used to contribute to energy conservation. For example, it can be used to optimize industrial processes to improve material and energy utilization, optimize logistics to reduce vehicle unloading rates, optimize urban lighting and traffic, optimize the use of air conditioning and lighting in buildings according to people's work and rest time, and intelligent remote work platforms can help reduce unnecessary energy cost.

Contribute to reducing greenhouse gas emissions

AI can be used to contribute to reduce greenhouse gas emissions through process optimization and monitoring technology, and help develop new materials and new processes that release less greenhouse gases. AI can be used to optimize breeds and agricultural production methods to reduce methane and nitrogen oxides in animal husbandry and agriculture. AI can be used for automatic garbage classification to recycle and obtain fuel, feed, steel and other materials, industrial raw materials, etc. in the garbage for reuse.

Promote greenhouse gas absorption and carbon storage

AI can be used to monitor, predict the greenhouse gas production from the ecosystems, and contribute to the intervention process. AI can be used to promote the development of greenhouse gas recycling technologies, and to protect ecosystems such as forests, peatlands, and oceans that can absorb and store carbon.

Reducing the harm caused by climate change

AI can be used to monitor, simulate, and predict extreme climates to reduce their harm, and can be used to make rational planning for agriculture under global warming and assist in the development of crops adapted to the environment change to reduce the impact on agriculture. AI should assist urban planning so that it can cope with the impact of climate change. AI should assist in optimizing industrial processes, enabling companies to comply with green standards for sustainable development.

Empower the development of energy systems

AI should contribute to the development of the energy system. AI can be used to optimize energy systems and promote green energy development. For example, AI can be used to optimize the overall design and deployment of energy systems, and improve energy efficiency. AI can also contribute to research, development and use of green energy technologies such as wind power, photovoltaics, nuclear fusion, geothermal energy, biogas, and electric fuel. Through the development of smart grid control and related technologies, AI can be used to improve the stability and reliability of the power systems.

Contribute to the establishment of market mechanisms and policies conducive to the control of climate change

AI can be used to analyze and predict the carbon footprint of various industries, to help to establish a scientific and effective carbon trading mechanism and a real-time energy pricing system, so as to promote the establishment of a market mechanism conducive to climate change control. With careful ethics evaluation, sufficient investment and sound infrastructure should be given to AI-related projects that are conducive to climate change control.

Partner Institutions

International Research Center for AI Ethics and Governance, Institute of Automation, Chinese Academy of Sciences [Lead drafting and organizing partner]

Institute for AI International Governance, Tsinghua University

Institute for Ethics in Artificial Intelligence, Technical University of Munich

Center for AI and Data Governance, Singapore Management University

SPARK UNDP Sustainable Development Goals Innovation Lab (Chengdu)

Stanley Ho Big Data Decision Analytics Research Centre, The Chinese University of Hong Kong (CUHK)

Centre for Perceptual and Interactive Intelligence, a CUHK InnoCentre

AI for Sustainable Development Goals Cooperation Network

Centre for AI Research (South Africa)

Research Center for Brain-inspired Intelligence, Institute of Automation, Chinese Academy of Sciences

China Institute for Science and Technology Policy, Tsinghua University

Article Link: https://ai-for-sdgs.academy/principles-on-ai-for-climate-action

上一条:Ethics As We Know it is Gone. It's Time for Ethics Re-envisioned