Climate Cost of AI
This topic of “Climate Cost of AI” is important from the perspective of the UPSC IAS Examination, which falls under General Studies Portion.
Background:
- Techno-optimism is an ideology that believes that technological development will help in the betterment of the society. Artificial Intelligence is one such technology that is considered as a gateway to the future.
- In the recent budget speech, the Finance Minister described the technology as ‘a sunrise sector’ that would “assist sustainable development at scale and modernise the country”.
How fair is the race for AI dominance?
Resources:
- The race for dominance in the sector isn’t fair- only a few developed countries possess the material advantages needed for the sector to progress, right from the start.
- These countries have the advantage of possessing both cutting edge R&D and the skilled human resource, in addition to sufficient wealth to enable investment in this technology.
- For instance, North America and East Asia account for 3/4th of the total private investment in AI related patents and publication.
- Consequently, these few economies also set the rules for the sector.
Governance:
- There is also inequity in terms of governance i.e. developing and underdeveloped nations face challenges in terms of:
- ‘Tech fluency’ of policymakers
- Barriers in formulation of regulations and industrial policy
- Insufficient representation and lack of empowerment at international forums that set AI-related rules and standards
What is the climate cost of AI?
- AI models are trained before being deployed to perform their myriad functions. This means that the model is taught to interpret data properly and accurately perform its function.
- For this training process and for the model’s operation, a lot of energy is consumed.
- In 2020, digital technologies accounted for 1.8% to 6.3% of global emissions.
- Meanwhile, AI development and deployment has increased manifold across the world. Also, the demand for processing power associated with increasingly larger AI models is increasing.
- In addition to this, governments of the Global South consider the technology to be a silver bullet for addressing complex socio-economic problems.
- From this, it can be concluded that we could witness a growing share of AI in digital technologies-related emissions in the coming years.
What has been done to address the issue?
- The concept of sustainability is increasingly featuring in mainstream debates around ethics and sustainable development of AI.
- In November 2021, the Recommendation on the Ethics of Artificial Intelligence was adopted by UNESCO. It calls upon stakeholders to “reduce the environmental impact of AI systems, including but not limited to its carbon footprint.”
- Tech giants like Microsoft, Facebook, Apple and Amazon have rolled out ‘net zero’ policies and initiatives.
What is the way ahead?
- These initiatives are a good sign. However, they only scratch the surface of the issue. This is especially so because global governance of AI and climate policy are contentious issues and rooted in inequitable access to resources.
- The Global South faces challenges on 2 fronts:
-
- The social and economic benefits of the technology are accruing to only a few countries.
- Most of the efforts regarding and narratives on the climate impact of AI are being dictated by the Global North.
- Also, like other nexus issues, the relationship between AI and climate change is still an understudied concept. This is because companies working in their field aren’t transparent or are not meaningfully committed to even studying, let alone acting to solve the issue.
- Developing countries need to assess their tech-led growth priorities while keeping the climate cost of AI in mind.
- As developing countries aren’t weighed down by legacy infrastructure, they would find it easier to ‘build up better’, i.e. they don’t have to copy what the West did with their AI-led growth paradigm. They could chart more sustainable paths.
- These countries should consider solutions that would truly cater to their unique social and economic requirements.
Conclusion:
National dreams of economic development and improved competitiveness, underwritten by AI, do have an allure. However, there are environmental costs and there is also the possibility of getting locked into rules set by powerful countries.
Referred Sources
