Philip Smith, journalist

With the news that fewer than half of all professional accountants have a basic understanding of how an artificial intelligence (AI) algorithm works, now would appear to be the time to invest in building awareness of the impact that AI can have on ethics and sustainability.

This is just one finding revealed in a wide-ranging study of the impact that the rapid adoption of AI is having on organisations around the world. In particular, the study, Ethics for sustainable AI adoption (produced in partnership with CA ANZ with whom ACCA has a Strategic Alliance), focuses on the importance of prioritising an ethical approach to using AI and how this cuts across all three of the environmental, social and governance (ESG) aspects of running an organisation.

'The ethical deployment of AI will need checks and balances – an area where accountants can lead'

This is particularly important for professional accountants; the survey reveals that one in five (19%) of accountants say that their organisation has adopted AI within finance-related tasks or functions, with 7% saying that it is already included in audit and assurance work.

However, at the same time, the survey also reveals that more than half (51%) believe that AI is having a positive impact on their ability to live according to their values.

The study reports that the past few years have seen intense global activity to establish some essential principles underlying an ethical approach to AI. While there are differences in emphasis, they tend to agree on certain broad principles for a responsible AI system. These include fairness, accountability, sustainability and transparency, as well as human oversight, the ethical use of data, safety and robustness, and standards and law.

Data: the carbon cost

Storing and using data comes with a cost – when data is stored in ‘the cloud’, this really means that it is physically stored on a server, or a network of servers firmly placed on the ground. We are all familiar with megabytes and gigabytes, perhaps terabytes and petabytes as terms describing data storage sizes. But what about zettabytes (1021 bytes) or yottabytes (1,000 zettabytes)?

An estimated 59 ZB of data was created around the world in 2020. As the absolute amount of data reaches previously unseen levels, the rate at which this new data is added is also greater than ever. The volume of data is massive, and it is growing exponentially, not linearly.

The servers that store this mountain of data need to be powered. As such, an AI system reliant on this data will have an identifiable carbon footprint. And it is not trivial.

But paradoxically AI can play a role in helping reduce its own carbon footprint. Multi-dimensional real-time interactions between data centre equipment, cloud infrastructure, cooling systems, electricity generators and human operators can be modelled using machine learning.

So, while the power consumed by AI algorithms can inhibit work towards achieving many of the UN’s sustainable development goals, such as carbon emissions, it can help achieve many more.

Accountants centre stage

As such, professional accountants are well-placed to take a lead on these ethical principles. ‘Finance leaders have an opportunity to leverage their sound ethical judgment alongside commercial and operational knowledge,’ the author of the report, ACCA’s head of business insights Narayanan Vaidyanathan, says. ‘The ethical deployment of AI will need checks and balances to ensure long-term value – an area where they can lead.’

Participants in the survey agree. Karen Smith FCCA, a partner in IBM, says: ‘Finance leaders have a mix of strategic, financial, operational and governance skills that make them ideal for driving the adoption of ethical practices when using AI in their organisations.’

This is why understanding the basics of any algorithm is so important. Algorithms are shaped by ideas, cultures and values. This, therefore, places an onus on an accountant’s professional competence and due care to understand what the algorithm is doing.

It also requires integrity in not passing accountability to the algorithm itself; professional judgment cannot be replaced by a compliance-based checklist, the report’s authors argue.

All in the data

Data lies at the heart of AI. The survey results reveal that about half the respondents (51%) consider their organisations to be effective at maintaining both data quality and data confidentiality. But while 25% report that their organisations are ‘very effective’ at managing confidentiality, only 16% claim this level of effectiveness for data quality. The difference could be linked to the compliance aspect of the former, the report concludes.

While there is some similarity in the overall picture of the level of effectiveness across quality and confidentiality, the two differ on where the challenges to effectiveness lie. For data quality, the core issue is the point when the data first enters the organisation. But for confidentiality, the biggest area of challenge reported was in the secure storage phase of the data lifecycle. See graphics below.

‘AI is one of the most exciting, transformational technological developments of our time,’ Vaidyanathan says. ‘But technology has the potential both to improve lives and to cause harm. Ultimately, it is the ethical and sustainable adoption of AI that will determine its relevance and usability.’