If you work on public procurement and your inbox looks anything like mine after the summer, you probably have more than a few emails pitching what AI can do for you, offering AI-based ‘hacks’ for procurement professionals, or reporting on new developments in AI and digital regulation that can be of relevance in the context of procurement. You may have also seen recent think tank reports, such as this one, on how to reimagine procurement for the AI era. And there is an emerging field of AI procurement-based humour, if that is your cup of tea.

Chances are that, even if your organisation is not considering using AI to manage procurement, it is likely to be exploring how AI can be used in its frontline activities. Sooner or later, if they haven’t already, most public buyers will be asked to explore ways in which to safely procure AI-based solutions, or will be offered such solutions within the tenders they receive, as providers are also quickly exploring ways to embed AI as part of their delivery processes.

In short, AI has become a focus of attention for procurement professionals and researchers.

However, as I discovered while carrying out research for my book on digital technologies and public procurement, it can be hard to find good resources to cut through the hype and understand what AI is, how it works (or doesn’t), and what implications and risks it carries. There is also little discussion of the (boring, expensive) preparatory work required to successfully deploy AI, and on the safeguards that need to be put in place.

The truth is, however, that ‘adopting AI’ cannot be done overnight and that organisations that rely on fully outsourced ‘AI as a service’ solutions, or that inadvertently embed elements of AI—for example, through cloud-based services they already have in place—are opening themselves up to cyber and commercial risks that are not always well understood or properly addressed. There have been some high-profile embarrassing cases of AI going rogue, such as in Air Canada’s chatbot fiasco, and it is becoming increasingly clear that ‘rushing to AI’ has both operational and reputational risks. AI-related decisions thus require careful consideration.

In the public sector context, procurement officials will be at the forefront of such decisions.

The question is how well equipped we (as a profession) are to make those calls, and how to improve our ability to do so. Although there are some sources of guidance—for example, in the UK the Guidelines for AI Procurement, the Guide to Using AI in the Public Sector, or the more recent Introduction to AI Assurance—they are not necessarily ‘plug and play’ and procurement professionals are expected to be able to exercise judgment in relation to increasingly complex issues surrounding AI development and deployment.

This will not change under the new legislation that is quickly being put in place to create a system that promotes trustworthy AI. Although the EU AI Act was adopted in July, it will not fully enter into force for another two years and, even then, its operationalisation through procurement documents and contractual clauses will require some careful thinking.

The successful deployment of AI across the public sector, and within public procurement, will thus crucially depend on the understanding and digital skills of procurement professionals. In my view, it will not be enough for some ‘digital specialists’ to be brought into procurement teams. All procurement professionals will need to have a good understanding of key issues. So this is a good time to start developing or increasing your level of digital understanding.