Artificial intelligence (AI) is at the heart of transformation and offers immense potential for innovation, and the capability to address the most pressing challenges facing the continent and the world.
So said Professor Tshilidzi Marwala, rector of the United Nations University and under-secretary general of the United Nations, speaking today at the 17th International Conference on Theory and Practice of Electronic Governance (ICEGov).
Marwala unpacked the challenges and opportunities inherent within AI, particularly with regards to its responsible and realistic governance.
“As we embrace these digital frontiers, we must confront a pressing question – how do we govern this field responsibly and ethically? When we talk about governing AI, what are we actually governing?”
According to Marwala, it comes down to four factors – data, algorithms, computing, and the applications of this technology across sectors and industry.
These factors form the foundations of governance conversations that must then be further shaped by two principal documents – the Declaration of Human Rights and the UN Charter, as they offer insight and guidelines into issues around transparency, security and explainability, he noted.
Marwala believes this foundation is the first step towards responsible and relevant AI governance.
“The second issue is to examine governance around human and organisational behaviour, as this technology is being used by people, and used and made by organisations,” he added. “This makes the field of behavioural science critical in understanding the framework of governance within the human context.”
The third consideration is mechanism design – the use of mechanisms designed to incentivise and disincentivise governance, said the professor.
How can organisations and individuals be incentivised to shift their focus to prioritise development and use cases for AI within governance structures? These are the foundational questions that should be asked when it comes to creating governance structures for AI that have longevity and relevance, he commented.
However, Marwala is quick to highlight that morality and values cannot be legislated, so perhaps the question should be whether transparency should be legislated. Should governance prioritise which algorithms companies or individuals can use?
“If algorithmic choices are regulated, this becomes an AI balancing problem,” says Marwala. “We need to balance progress, while addressing the risks. Find a way of balancing global governance alongside local governance.
“I don’t believe the answer lies in the extremes, but in the optimal combination of all these factors,” said Marwala. “As we govern this technology, we need to do so within the context of the Sustainable Development Goals (SDGs), particularly in light of the recent scorecard which showed that only 17% of 179 indicators of the SDGs are being met. The rest are worse off than 2015 or lagging behind.”
A recent study undertaken by Ricardo Vinuesa, KTH Royal Institute of Technology, found that AI and machine learning have immense potential to support the SDGs, potentially enabling 134 of the 169 SDGs. The technology also has the potential to inhibit 59 of these goals.
“As we look to AI to support the 2030 agenda, we must recognise how it can both enhance our ability to combat challenges and the inequalities we seek to eradicate,” said Marwala.
“This year alone, more than two billion people participated in democratic elections globally and yet deep fakes are undermining the integrity of the democratic process and are a threat to global democracy.
“Addressing bias is not just technical, it’s a moral imperative,” said Marwala. “Internationally, there is progress as the UN member states have established a pact to promote sustainable development through collaboration and innovation. It is a global digital compact that can provide a guiding framework going forward.
“The biggest problem we face as international organisations and government is we talk more than we do. The time of action is now and we cannot afford to let the time lapse as the consequences will be damaging.”
Share