Environmental and societal well-being¶

The EU Ethical AI Guidelines notes the following on this topic:

“AI systems should benefit all human beings, including future generations. It must hence be ensured that they are sustainable and environmentally friendly. Moreover, they should take into account the environment, including other living beings, and their social and societal impact should be carefully considered.”

This guideline, although a bit philosophical, means that we should be raising awareness of the impact of AI on the evironment and societal well-being. The following table has the purpose to raise this awareness and it mentions some practical steps that AI contributors may take in this regard.

Topic	Guideline	Practical consequences and examples
Sustainable and environmentally friendly	AI contributors must design AI systems so that the resources used to implement them minimize the environment footprint, whether it is during the model development phase, model deployment phase, or during the inference phase. They should report environment-related metrics, and account for them as much as they account for pure machine learning or $-related performance metrics.	Estimate and report to stakeholders greenhouse gas emissions (GHG) , using ML CO2 impact or Green-Algorithms for example. Use compute- efficient ML: This allows for the use of lower-cost resources like CPUs compared to more expensive hardware for the training of complex architectures which typically require GPUs or TPUs. A comparison of CPUs, GPUs and TPUs in regards to their cost-energy-performance, can be found here. Use federated learning: This enables computations to run locally thus potentially decreases carbon impacts if the computations are done in a place where electricity is generated using clean sources. Data storage and usability: While JSON is more efficient in terms of compute storage and perhaps memory, therefore environmentally efficient; conversion of a JSON file to CSV/EXCEL is very costly and significantly increases environmental footprint. Take collaboration/sharing of files into account when deciding on file type(s).
Social, societal	AI systems may alter our conception of social agency or impact our social relationships. While AI systems may enhance social skills, it might equally contribute to social deterioration, affecting people’s physical and mental wellbeing. The effects an AI system might have on this should be carefully tracked and monitored.	AI impacts the following societal topics: Economics: AI technologies bring an industrial revolution. However, AI also increases wealth inequality. Public Health: AI is rapidly evolving in the field of healthcare. This might have consequences for human wellbeing and safety. Labor market: AI takes over tasks previously done by human workers. This might lead to unemployment. However, this might also create (new) jobs and specializations. Security: AI systems might be hacked by cybercriminals, allowing them to do harm to society.
Democracy	The effects of an AI system on institutions, democracy and society at large should be taken into account. In particular, situations related to the democratic process, including not only political decision making but also in electoral context.	Use federated learning (decentralized data) : Federated learning mitigates risk and harm that arises from centralized data, including data breaches and privacy intrusions. Impose data sovereignty: individuals should be given control over how their data is used, for what purposes and how long. It allows users to withdraw consent if they see fit. Some well-known incidents exist where data and AI is/was wrongfully used to influence society/democracy and breach people’s privacy: Cambridge Analytica – Facebook scandal Clearview Face Recognition scandal Mass surveillance without consent Deepfakes

Topic

Guideline

Practical consequences and examples

Sustainable and environmentally friendly

AI contributors must design AI systems so that the resources used to implement them minimize the environment footprint, whether it is during the model development phase, model deployment phase, or during the inference phase. They should report environment-related metrics, and account for them as much as they account for pure machine learning or $-related performance metrics.

Estimate and report to stakeholders greenhouse gas emissions (GHG) , using ML CO2 impact or Green-Algorithms for example.
Use compute- efficient ML: This allows for the use of lower-cost resources like CPUs compared to more expensive hardware for the training of complex architectures which typically require GPUs or TPUs. A comparison of CPUs, GPUs and TPUs in regards to their cost-energy-performance, can be found here.
Use federated learning: This enables computations to run locally thus potentially decreases carbon impacts if the computations are done in a place where electricity is generated using clean sources.
Data storage and usability: While JSON is more efficient in terms of compute storage and perhaps memory, therefore environmentally efficient; conversion of a JSON file to CSV/EXCEL is very costly and significantly increases environmental footprint. Take collaboration/sharing of files into account when deciding on file type(s).

Social, societal

AI systems may alter our conception of social agency or impact our social relationships. While AI systems may enhance social skills, it might equally contribute to social deterioration, affecting people’s physical and mental wellbeing. The effects an AI system might have on this should be carefully tracked and monitored.

AI impacts the following societal topics:
- Economics: AI technologies bring an industrial revolution. However, AI also increases wealth inequality.
- Public Health: AI is rapidly evolving in the field of healthcare. This might have consequences for human wellbeing and safety.
- Labor market: AI takes over tasks previously done by human workers. This might lead to unemployment. However, this might also create (new) jobs and specializations.
- Security: AI systems might be hacked by cybercriminals, allowing them to do harm to society.

Democracy

The effects of an AI system on institutions, democracy and society at large should be taken into account. In particular, situations related to the democratic process, including not only political decision making but also in electoral context.

Use federated learning (decentralized data) : Federated learning mitigates risk and harm that arises from centralized data, including data breaches and privacy intrusions.
Impose data sovereignty: individuals should be given control over how their data is used, for what purposes and how long. It allows users to withdraw consent if they see fit.
Some well-known incidents exist where data and AI is/was wrongfully used to influence society/democracy and breach people’s privacy:

Appendix - Recommendations from the EU¶

Below are the recommendations directly reported from EU.