Ethical AI: Preparing your organisation for the future of AI

By Rosemary J Thomas, Senior Technical Researcher, AI Labs Version 1

Artificial intelligence is changing the world, generating countless new opportunities for organisations and individuals. Conversely, it also poses several known ethical and safety risks, such as bias, discrimination, privacy violations, alongside its potential to negatively impact society, well-being, and nature. It is therefore fundamental that this groundbreaking technology is approached with an ethical mindset, adapting practices to make sure it is used in a responsible, trustworthy, and beneficial way.

To achieve this, first we need to understand what an ethical AI mindset is, why it needs to be central, and how we can establish ethical principles and direct behavioural changes across an organisation. We must then develop a plan to steer ethical AI from within and be prepared to take liability for the outcomes of any AI system.

What is an ethical AI mindset

An ethical AI mindset is one that acknowledges the technology’s influence on people, society, and the world, and understands its potential consequences. It is based on the perception that AI is a dominant force that can sculpt the future of humankind. An ethical AI mindset ensures AI is allied with human principles and goals, and that it is used to support the common good and the ethical development of all.

It is not only about preventing or moderating the adverse effects of AI, but also about exploiting its immense capability and prospects. This includes developing and employing AI systems that are ethical, safe, fair, transparent, responsible, and inclusive, and that respect human values, autonomy, and diversity. It also means ensuring that AI is open, reasonably priced, and useful for everyone – especially the most susceptible and marginalised clusters in our society.

Why you need an ethical AI mindset

Functioning with an ethical AI mindset is essential.  Not only because it is the right thing to do, but also because it is expected, with research showing customers are far less likely to buy from unethical establishments. As AI evolves, the expectation for businesses to use it responsibly will continue to grow.

Adopting an ethical AI mindset can also help in adhering to current, and continuously developing, regulation and guidelines. Governing bodies around the world are establishing numerous frameworks and standards to make sure AI is used in an ethical and safe way and, by creating an ethical AI mindset, we can ensure AI systems meet these requirements, and prevent any prospective fines, penalties, or court cases.

Additionally, the right mindset will promote the development of AI systems that are more helpful, competent, and pioneering. By studying the ethical and social dimensions of AI, we can invent systems that are more aligned with the needs, choices, and principles of our customers and stakeholders, and can provide moral solutions and enhanced user experiences.

Ethical AI as the business differentiator

Fostering an ethical AI mindset is not a matter of singular choice or accountability, it is a united, organisational undertaking. To integrate an ethical culture and steer behavioural changes across the business, we need to take a universal and methodical approach.

It is important that the entire workforce, including executives and leadership, are educated on the need for AI ethics and its use as a business differentiator. To achieve this, consider taking a mixed approach to increase awareness across the company, using mediums such as webinars, newsletters, podcasts, blogs, or social media. For example, your company website can be used to share significant examples, case studies, best practices, and lessons learned from around the globe where AI practices have effectively been implemented. In addition, guest sessions with researchers, consultants, or even collaborations with academic research institutions can help to communicate insights and guidance on AI ethics and showcase it as a business differentiator.

It is also essential to take responsibility for the consequences of any AI system that is developed for practical applications, despite where organisations or products sits in the value chain. This will help build credibility and transparency with stakeholders, customers, and the public.

Evaluating ethics in AI

We cannot monitor or manage what we cannot review, which is why we must establish a method of evaluating ethics in AI. There are a number of tools and systems than can be used to steer ethical AI, which can be supported by ethical AI frameworks, authority structures and the Ethics Canvas.

An ethical AI framework is a group of values and principles that acts as a handbook for your organisation’s use of AI. This can be adopted, adapted, or built to suit your organisation’s own goals and values, with the stakeholders involved in its creation. An example of this can be seen in the UK Government’s Ethical AI Framework, and the Information Commissioner’s Office’s AI and data protection risk toolkit which covers all ethical risks in the lifecycle stages – from business requirements and design to deployment and monitoring for AI systems.

An ethical AI authority structure is a group of roles, obligations and methods that make sure your ethical AI framework is followed and reviewed. You can establish an ethical AI authority structure that covers several aspects and degrees of your organisation and delegates clear obligations to each stakeholder.

The Ethics Canvas can be used in AI engagements to help build AI systems with ethics integrated into development. It helps teams identify potential ethical issues that could arise from the use of AI and develop guidelines to avoid them. It also promotes transparency by providing clear explanations of how the technology works and how decisions are made and can further increase stakeholder engagement to gather input and feedback on the ethical aspects of the AI project. This canvas helps to structure risk assessment and can serve as a communication tool to convey the organisation’s commitment to ethical AI practices.

Ethical AI implications

Any innovation process, whether it involves AI or not, can be marred a fear of failure and the desire to be successful in the first attempt. But failures should be regarded as lessons and used to improve ethical experiences in AI.

To ensure AI is being used responsibly, we need to identify what ethics means in the context of our business operations. Once this has been established, we can personalise our message to the target stakeholders, staying within our own definition of ethics and including the use of AI within our organisation’s wider purpose, mission, and vision.

In doing so, we can draw more attention towards the need for responsible use policies and an ethical approach to AI, which will be increasingly important as the capabilities of AI evolve, and its prevalence within businesses continues to grow.