The development of AI is influencing a growing number of industries. It is anticipated that AI will have immediate and long-term effects on various factors, including but not limited to global productivity, equality and inclusion, environmental impacts, and others. 

Both positive and negative influences on sustainable development have been reported as possible outcomes of AI. As we find that AI has the potential to affect the ability to accomplish the SDGs, this is a crucial area for future study.

A team of Swedish academics presents and discusses the implications of artificial intelligence's potential to aid or impede the achievement of each of the 17 goals and 169 targets outlined in the 2030 Agenda for Sustainable Development. The methodology provided at the study's conclusion can be summed up as a consensus-based expert elicitation approach, guided by earlier studies aiming at mapping SDGs interlinkages. 

For this research, although there is no universally accepted definition of artificial intelligence, the researchers included AI, any software technology with at least one of the following capabilities:

• Perception (e.g., face recognition).
• Decision (e.g., medical diagnosis systems).
• Prediction (e.g., weather forecast).
• Automatic knowledge extraction and 
• Pattern recognition from data (e.g., the discovery of fake news circles in social media).
• Interactive computational complexity (e.g., natural language (e.g., theory development from premises).

Among the many specialised areas that this perspective incorporates is machine learning.

AI's impact on society

Among the social group in the figure below, 67 objectives (82%) might benefit from AI-based technology. For instance, AI can enable smart and low-carbon cities comprising a variety of related technologies, such as autonomous electric vehicles and smart appliances, which can facilitate demand response in the electricity sector.

Image source: Nature

By enabling smart grids that partially match energy demand to periods when the sun is shining, and the wind is blowing, AI can also help to integrate variable renewables. 

Advanced AI technologies, research, and product creation may necessitate vast computer resources only available in substantial computing centres. These facilities have a very high energy demand and carbon footprint. For instance, cryptocurrency applications such as Bitcoin consume as much electricity globally as the electrical demand of some nations, jeopardising not only SDG 7 but also SDG 13 on Climate Action. According to some estimates, the total electricity demand for information and communications technologies (ICTs) might account for up to 20% of the worldwide electricity demand by 2030, up from approximately 1%. The development of ICT technologies in a sustainable manner is therefore vital. Energy-efficient cooling solutions for data centres, greater energy efficiency, and using renewable energy in ICTs will all contribute to regulating electricity consumption growth.

Conclusion

In addition to more efficient data centres powered by renewable energy, it is vital to incorporate human knowledge into constructing AI models. In addition to the fact that the human brain consumes significantly less energy than is required to train AI models, the available knowledge introduced in the model (see, for example, physics-informed deep learning) does not need to be learned through data-intensive training, which can significantly reduce the associated energy consumption. 

Although AI-enabled technology can serve as a driver for achieving the 2030 Agenda, it may also result in disparities that hinder the achievement of SDGs 1, 4, and 5. On the other hand, it may also lead to additional qualification requirements for all jobs, aggravating existing inequities and impeding the fulfilment of this objective.