Algorithms: what every auditor needs to know

Our organisations are increasingly shaped by algorithms; the unseen influence in our lives. Yet how often are they mentioned in the scope of an audit engagement or a final report?

Don’t click away…..this is not a technical read! It’s for all audit leaders.

Algorithms are part of the digital workforce, an unseen labour force, increasing efficiency by rapidly assimilating and assessing data. But can we trust them? How well do we know them? Who recruited them? What vetting did they go through?

What risks is your organisation exposed to by the algorithms that are used?
Do you know? Does anyone in your organisation know?


What do we mean by algorithms?

Algorithms like most things computer orientated are surrounded by mystery, a black box of technical complexity, the realm of computer programmers or nerdy Excel wizards; but in reality they are not complicated. They are quite simply a set of instructions, the maths that enable computers to solve a problem or complete a task; control driverless cars, compose music or decide which adverts are seen on your social media. Tool such as decision trees, statistical models, neural nets, regression techniques and Bayesian models are all algorithms.

In our age of big data, they are also incredibly powerful control mechanisms deciding what information we are shown on internet search engines, screening job applicants and determining credit ratings. The majority of stock market movements around the world are via algorithm trading, the Internet of Things is powered by algorithms and supply chain logistics are being revolutionised by machine learning (programming that enables computers to adapt and learn from ongoing data analysis rather than follow one set of rules).

It’s not uncommon to hear IT auditors say that they will audit up to and around the ‘box’ but not through it, meaning that the…