It is clear that a lack of trust in AI outputs is an issue for a number of respondents. However, these reservations could be more of a reflection of the lack of understanding of how these systems work. An alternative view is that AI and machine learning can potentially increase the credibility and accuracy of insights rather than detract from them. This rigor is due to the fact they arrive at conclusions based on a larger number of data sets, rather than an individual probing a single set of data and potentially introducing their own biases into the equation. It is therefore likely that smart machines could undertake data-driven tasks with greater accuracy, consistency and time-efficiency than humans.
So the question becomes: how can finance functions build trust into the AI they use to drive insight? Without it, stakeholder confidence in the AI is likely to prove elusive. The starting point is for finance leaders to understand and map some of the new risks that AI brings and to use these insights to begin to create the right governance and control mechanisms. To build trust in the outputs, AI should be trained properly, with appropriate boundaries around it at first. Then, after it is put into production to identify and rectify any flaws, its performance should be continually monitored. It’s another new demand of finance leaders, whose evolving responsibilities are examined in this CFO Imperative series, which identifies critical answers and actions to help leaders reframe the future of their organizations.
Rethinking the future finance talent strategy
The impact of AI could also be profound for the people in the finance team and how CFOs think about the future talent strategy. In the survey, 64% of respondents said a wide range of core finance roles – such as financial reporting, accounting and financial control – could be significantly disrupted and changed as a result of advances in automation and AI. The function’s skills profile is also likely to change dramatically in the future to be more digitally focused. As shown below, cognitive computing skills could be the most in-demand, followed by technology delivery and digital transformation. And as shown below, respondents said that cognitive computing is the most important skill required both over the next 12 months and the next 3 years.