Chief Managing Editor,
The Indian Learning.
On the 8th of April 2020, the Committee of Ministers of the Council of Europe released a recommendations report that details extensively on the potential impact of the use of algorithms on human rights protected for in the European Convention on Human Rights, 1953. This report is inclusive, broad and holistic in its narrative, and covers a wide range of possibilities and scenarios wherein regular use of algorithms can hinder the regime of European Human Rights Law.
The objective for this report is laid out in its beginning; citing the factors of the unprecedented use of digital applications across various sectors of modern life, and the evolving use of algorithmic systems, the impact and scope of use of which is still not fully comprehended, as two critical reasons behind the urgency of the need for a body of ethical standards on the use of algorithms. This is to ensure that the use of algorithmic systems doesn’t further fortify existing biases and discriminations which are still prevalent in the society that the body of Human Rights Law aims to curb and regulate through its enforcement. In their efforts to elaborate on the human rights impact of the use of these systems, the Committee of Ministers recognises that public and private sector initiatives are already underway towards the manifestation of bodies of ethical standards and regulations, but such attempts in no manner relieve the Council of Europe of its duty to protect the convention.
The beginning set of recommendations requests all the member states to conduct the following actions with respect to the guidelines presented later in the report, namely:
To review their legislation, frameworks, policies, and their own practices with respect to the procurement and development of algorithmic systems,
Create effective legislations that govern over private-sector effort towards the development of the algorithmic system to ensure their compliance with the international regime,
Engage in regular and transparent consultations and dialogue with all involved stakeholders and members of their society,
Prioritise the inculcation of public and private expertise in the subject area,
Promote the importance and necessity of digital literacy,
And take into account any environmental impact of large-scale digital services, and strive towards creating sustainable models with optimised use of energy and natural resources.
As per the 2nd guideline under section (A) of the report, the body recognises algorithmic systems as
“applications that, often using mathematical optimisation techniques, perform one or more tasks such as gathering, combining, cleaning, sorting, classifying and inferring data, as well as selection, prioritisation, the making of recommendations and decision making. Relying on one or more algorithms to fulfil their requirements in the settings in which they are applied, algorithmic systems automate activities in a way that allows the creation of adaptive services at scale and in real time”.
This definition is very broad, inclusive and based in the rational-centric approach to AI, which detaches itself from a human-centric approach to assessing AI. This also essentially “future-proof” the scope of the definition’s application to any future developments we could witness in AI research & development, with any potential form of algorithmic systems, while likely to have unique traits and functions, fitting within the scope of this definition as it would still include all the elements as mentioned in it.
While Section A of the guidelines continues into explaining the intricacy, utility and potential issues that could arise out of regular use of algorithmic systems, Sections B & C of the guidelines effectively chart out recommendations towards member States and the private sectors within those states respectively. Alongside a common set of recommendations towards both recipients, namely the management of effective systems of data that harbour accurate analyses and modelling capabilities, the protection and observation of human rights, protection and respect of personal data and the maintenance of accurate representations of data, both parties are recommended a set of guidelines to observe independent of each other. States are recommended to make any legislation concerning algorithmic systems transparent and accountable, whilst upholding the principles of democracy and ensuring the consistent review of these systems to ensure a minimal occurrence of error. The private sector, on the other hand, is recommended investing their efforts into developing systems that eliminate biases and discriminations that amount to violations of human rights and ensure transparency and accountability towards consumers in the creation of goods and services that utilise these systems.
The recommendations and guidelines listed out in this report are comprehensive, envisioned effective in their potential application and are truly progressive and one of a kind in the international regime on AI research and development. It serves as an effective catalyst between the establishment of human rights governance and 21st century technologically fuelled socio-economic development, ensuring that the order and protection guaranteed in the former don’t get lost in the chaotic and fast-paced developments of the latter. The report can serve as a guide for the drafting of effective legislations in all the European States, and many other states across the international regime, eventually creating a whole new domain of human rights law addressing the protection of it against the disruptive factors of the post-modern world.