Last Tuesday, the 10th of December was the annual international Human Rights Day. It’s a day which gives the global community an opportunity to focus on the importance of protecting the human rights of all citizens. Here in Scotland in a video message for the day I mentioned that it’s important on the international stage that we don’t lose sight of the work being put together to create a Convention on the Human Rights of Older Persons about which I wrote a few months ago. But I also reflected in my message on why it was important that Scotland should seek to re-introduce the Human Rights Bill which had earlier been proposed and around which I have already written. The decision of the current Scottish administration not to continue with the Human Rights Bill is very disappointing not least because in times of straightened budgets and real challenge in the world of social care the rights and dignity of older persons needs a special attention and focus, and a legal framework of rights enables that to happen with a sharper intensity.

More broadly, however, the issue of human rights in social care and aged care in particular has been uppermost in my mind. I’ve been very honoured to have chaired a small working group of individuals who have been concerned about the developing role of artificial intelligence in the delivery of aged care. The group has been working under the auspices of the Oxford Institute for Ethics in AI and the Digital Care Hub. It has developed a set of principles and guidance together with case studies to describe what we considered to be the key components which should lie behind and influence the use of artificial intelligence models and tools in the delivery of Adult Social Care.

This I think is really important because it is in the interface of technology and social care that there can be a danger that the rights and dignity of older persons or indeed any other group of people can be at risk. This is especially, I think, the case in these current times when decisions might be made on economic grounds rather than on the basis of the best interests of individual citizens.

So, I want in what remains of this blog to explore these principles but also to examine what is happening in the wider world of Ai around the practice of ethics and human rights not least in the various codes and conduct documents which have been developed to reassure those outside the tech industry that there are appropriate balances and safeguards which are there to protect us. Whilst many of these are indicative of a robust approach, I want to suggest that for those of us who operate in the world of social care that we need to increasingly be watchful and therefore the Oxford ethical approach and model for social care is all the more important.

The first thing to note is that it is certainly true that for bodies such as the United Nations and leading tech firms the establishment of frameworks to guide the responsible use of Ai has become an issue of real importance in the last couple of years. Take the United Nations for instance. In September 2022 the UN System Chief Executive Board for Coordination endorsed what they called the Principles for the Ethical Use of Artificial Intelligence in the United Nations System. In essence these ten principles emphasise the following:

• ‘Do No Harm: Ensuring Ai applications do not cause harm to individuals or communities.
• Defined Purpose, Necessity and Proportionality: Utilising Ai for clear, legitimate purposes aligned with UN values.
• Safety and Security: Safety and security risks should be identified, addressed and mitigated
• Fairness and Non-Discrimination: Preventing biases and ensuring equitable Ai outcomes.
• Sustainability: Artificial intelligence should be aimed at promoting environmental, economic and social sustainability
• Transparency and Explainability: Maintaining openness about Ai systems and their decision-making processes.
• Responsibility and Accountability: Assigning clear accountability for Ai operations.
• Privacy and Data Protection: Safeguarding personal data within Ai applications.
• Human Autonomy: Ensuring Ai supports human decision-making without undermining autonomy.
• Inclusion and participation: Organisations should take an inclusive, interdisciplinary and participatory approach and promote gender equality.’

We have also witnessed similar frameworks being developed by the major tech companies such as Microsoft who adopted ethical Ai principles focusing on fairness, reliability, privacy, inclusivity, transparency, and accountability. Google established AI principles emphasising socially beneficial Ai, avoiding harmful applications, and incorporating privacy design and IBM have sought to implement Ai ethics guidelines prioritizing trust and transparency.

This flurry of ethical activity led to eight global tech companies in February 2024 committing to applying UNESCO’s Recommendation on the Ethics of AI, integrating these values into their Ai development processes.

The Oxford Institute for Ethics in AI outlines seven ethical principles for AI in social care: truth, transparency, equity, trust, accessibility, responsiveness and humanity. These align closely with both UN and tech companies’ frameworks:

Truth and Transparency: Correspond with the UN’s emphasis on transparency and explainability, and tech companies’ focus on open Ai operations.

Equity: Reflects the UN’s fairness and non-discrimination principle and tech firms’ stated commitments to unbiased Ai.

Trust: Parallels the UN’s responsibility and accountability, fostering confidence in Ai systems.

Accessibility: Aligns with inclusivity efforts, ensuring Ai benefits are widely distributed.

Humanity: Echoes the UN’s human autonomy principle, ensuring Ai supports rather than replaces human roles.

Despite these alignments, certain gaps persist which I think will pose particular challenge and priority for those of us who work in social care in the months and years ahead, specifically

Implementation and enforcement: While principles are established consistent implementation and enforcement across very different contexts remains a challenge. What will be really important for the Oxford Principles is that they are continually evaluated, tested and assurance is sought that they are being implemented in as robust a manner as they should be.

Secondly addressing bias. There needs to be an ongoing effort to identify and mitigate biases in the Ai systems in order to prevent discrimination. It is sadly the experience of many of us who work in and deliver aged care services that the unconscious and sometimes subconscious age discrimination bias which exists within wider society is evidenced in so many of the systems of support and models of intervention. We need to make sure that Ai systems aren’t just non-discriminatory in terms of other characteristics but that they are definitively age unbiased.

In addition, it is important that we prioritise global inclusivity ensuring that Ai governance includes diverse perspectives from the whole global community and not solely those nations often described as being ‘developed’.  We risk a real marginalisation of communities and nations in the ongoing development of Ai not least in aged care if we do not ensure global inclusivity.

And lastly, transparency. Operationalising transparency and complex Ai systems requires continuous refinement. We are some distance away from that degree of sophistication which would ensure the absence of bias, consistent application and the enforcement and robust adherence to ethical principles whether that be the ethical framework developed by the UN or by the major technology companies.

I would suggest that these all to some extent align themselves with the common ground which the Oxford Institute Principles for Social Care describe. It is positive that there is such a convergence, but that reality also highlights the need to engage in continuous dialogue between those who are developing Ai tools and models, those who commission, contract and purchase such tools and fundamentally those who are impacted in the way in which they live their lives as individuals who use and access social care support.

There is a real possibility that Ai can positively revolutionise the support of some of our most valued citizens but if Human Rights Day shows anything it shows us the urgent continuous need to be ever watchful to ensure the realisation of human rights for all across our shared world.

Donald Macaskill

Photo by Markus Spiske on Unsplash

Last Updated on 14th December 2024 by donald.macaskill