The potential of AI is exciting, with many organisations rushing to embrace this next generation technology. Along with well documented benefits, AI can also be used to support those with additional needs, to build a fully inclusive society, but there are downsides for people and the planet which are often overlooked.
To bring true value to society and the environment, it is important to look at the overall impact of AI. We should aim to maximise AI’s positive effects while minimising any negative impact. This means making AI ethical and sustainable, reducing its harmful impacts, and enhancing its benefits, ensuring that AI is for ethics and sustainability.
Scrutiny of AI
In most cases, AI is not created to reinforce harmful biases or harm the environment, but it can sometimes do both. For example, the National Association for the Advancement of Colored People (NAACP), the civil right organisation in the United States of America, has called on lawmakers to evaluate and regulate the use of predictive policing and AI within law enforcement. While these tools aim to help law enforcement, there is mounting evidence that they can increase racial biases as a result of the black-box problem. For this reason, predictive policing has been banned in the EU’s AI Act, of which it is believed a Labour Government would look to pass a similar version if they were to win the general election in July 2024.
AI can also be unsustainable. Digital technology, including AI, has a huge environmental impact, contributing 3-4% of greenhouse gas emissions, more than aviation. AI data centres’ energy use is set to double in two years, matching Japan’s consumption. AI uses four times as much energy as standard cloud applications which if not sourced sustainably can worsen the climate crisis. To reduce the negative impact of AI, organisations and individuals need to take responsibility to ensure that their use of AI is sustainable.
Best practice for ethical and sustainable AI
So, what does best practice look like?
- Ethically, it means eliminating the black-box problem, ensuring AI outcomes are explainable and foreseen, and the use of data reflects the principles of trust and transparency, similar to how explainable AI is being used being used in certain areas of healthcare.
- From a sustainability perspective, it means that sustainable practices become embedded, such as prioritising renewable energy for data centres or using less data in the creation of AI. For example, all of Sopra Steria’s internal data centres are carbon neutral and are sourced with renewable energy. This approach to AI should be clear and present from design to deployment, reducing AI’s negative impact on society and the planet.
- To truly benefit society, AI shouldn’t just focus on driving financial value. AI should support and achieve equity for society. AI tools such as ChatGPT can boost efficiencies and cut costs, but they can also have a broader impact. For instance, the AI assistant “signbot” by Sopra Steria, IBM and IVèS enables digital sign language communication for the deaf and hard of hearing. Furthermore, Sopra Steria’s use of AI is opening up new opportunities to run efficient and sustainable precision forestry in Norway, demonstrating that AI can also deliver positive sustainable outcomes.
By using AI for ethical and sustainable causes, we can maximise its positive impact on society.
How to get the most out of AI
To get the most out of AI, we need to boost its positive effects and minimise the negatives. The larger the gap between these two, the more value that we can achieve, and the greater this gap, the more we ensure an ethical and sustainable future. It’s important to remember that sometimes the harm and negatives from AI are so severe and unacceptable, that they will stop its development and use.
Organisations need to be aware of this when making decisions on AI and act accordingly. They should always comply with the law and strive to be as ethical and sustainable as possible. To achieve this, they should make considering and recording the wider societal and environmental impact part of their AI strategy.
Being better by design
AI needs to be ethical and sustainable by design, but what does that mean? AI should be ethical and sustainable while also being for ethical and sustainable outcomes. This can be achieved by:
- Governance structures within organisations need to identify and prevent unethical or unsustainable AI use. Sopra Steria has set up an AI Governance board within the organisation and have helped clients do the same. These boards review all potential uses of AI and are empowered to prioritise ethical and sustainable considerations over profit.
- Employees should be educated on AI ethics and sustainability. By gaining knowledge to help reduce AI risks and create value, organisations can accelerate innovation, promote differentiation, and elevate customer trust. For example, 50% of Responsible AI leaders report better products and services as a result of their responsible AI efforts. Improved literacy includes approaches such as consequence scanning to reveal unintended impacts of AI, ensuring AI’s positive effects are maximised and negative ones are minimised. By investigating, documenting and managing these impacts, organisations can derive the most value from AI, while maintaining ethical and sustainable practices. In September 2023 Sopra Steria launched a transformation programme called rAIse focusing on Artificial Intelligence. The series is aimed at educating employees on AI, with the most recent session for employees being on AI and Sustainability. We’ve also launched the Sopra Steria Next Skills Academy for our clients where we run a session on Digital Ethics and AI, and will run a campaign for our people on AI to raise awareness across the organisation.
- Organisations should have a clear mission statement relating to how they will use AI, and what value it will bring that includes ethics and sustainability. Sopra Steria has a set of group principles on AI that outlines best practices to secure the way the organisation uses AI tools, and therefore stakeholders are aware of the organisation’s intention of AI and can align their behaviour accordingly.
Originally posted here
Read More AI for good
