Introduction by Emma O’Connell
Embarking on a career in public procurement is both exciting and daunting, especially in today’s ever-shifting digital panorama. As a new public procurement research officer, I find myself surrounded by conversations about digitalisation and artificial intelligence (AI). The potential of these technologies to revolutionise public procurement is immense, but so is the complexity they introduce. Where does one even begin in understanding AI’s role in procurement? Which questions should be asked, and to whom? These are the questions I grapple with as I start my career.
The significance of AI in public procurement cannot be overstated. AI has the power to revolutionise the way public procurement operates – enhancing efficiency, transparency, and cost-effectiveness. By automating routine tasks, AI allows procurement professionals to focus on making strategic decisions. AI’s data analytics capabilities can uncover insights that lead to better procurement strategies, identifying savings opportunities, and mitigating risks. For someone starting out, the potential of AI is both thrilling and overwhelming.
At the same time, AI is something that needs to be procured, whether for its use in procurement or in the broader public sector context. How are public buyers to approach the procurement of AI? What are the key considerations to be taken into account? How can public buyers assess their organisations’ AI readiness? To what extent is the freshly minted EU AI Act relevant?
The Starting Point: Understanding the Landscape
The first challenge is figuring out where to start. Public procurement is inherently complex, governed by a myriad of rules and regulations designed to ensure fairness, transparency, and accountability. Introducing AI into the mix adds another layer of complexity. There are so many questions to consider: What are the regulatory implications of using AI in procurement? How do we ensure that AI systems are fair and unbiased? What AI technologies are available, and how do they fit into the procurement process? These are just a few of the questions that come to mind as I try to make sense of this new frontier.
The discourse around AI in public procurement is widespread but often lacks practical guidance on how to begin. It’s easy to get lost in the sea of buzzwords and theoretical discussions without getting a clear picture of practical steps. Thus, navigating this new digital terrain requires identifying the right resources and experts. Knowing who to ask and what to ask is a challenge in itself. It’s crucial to connect with knowledgeable professionals who have hands-on experience integrating AI into procurement processes, and who can share insights on best practices and potential pitfalls.
Developing a Critical Eye
The noise around AI often makes it hard to discern which trends and developments are truly impactful. Leading experts are calling out the way AI is sometimes peddled as ‘snake oil’ and stressing the importance of raising the digital literacy of those involved in assessing, procuring, and using AI. Even where ‘the tech is sound’, not every technological advancement is relevant or beneficial to the procurement process. Therefore, it’s essential to develop a critical eye and learn how to evaluate the usefulness and feasibility of different AI applications. This involves understanding the specific needs of the procurement function within the organisation and identifying AI solutions that align with those needs.
Furthermore, ethical and legal considerations are paramount when integrating AI into public procurement. As someone starting out in this field, it’s vital to recognise the importance of ethical and legal AI practices. This includes ensuring that AI systems are transparent and explainable so stakeholders can understand how decisions are made, but also that these complex technological systems comply with rules on data protection, cyber security and, to the extent possible, the future obligations under the EU AI Act. It also involves being vigilant about potential biases in AI algorithms that could lead to unfair outcomes. Learning about the ethical implications of AI is a crucial part of my journey, as it ensures that the technology is used responsibly and in a way that upholds the principles of public procurement.
Learning from Experts
Starting out in the public procurement realm, my journey into AI in public procurement is just beginning. The importance of AI is clear, but so are the challenges and uncertainties. By asking the right questions, seeking out expert guidance, and continuously learning, I aim to navigate this new frontier and help shape the future of public procurement. With the right knowledge and skills, we can harness the power of AI to create more efficient, transparent, and ethical procurement processes.
That’s why I am eager to learn from experts such as Albert, who have extensive experience in this field. Their insights can help me understand what aspects of AI are most important for procurement professionals to focus on how to effectively implement AI solutions. By posing the right questions and seeking guidance from knowledgeable sources, I hope to navigate this complex digital sphere and contribute to the advancement of AI in public procurement.
Taking A Step Back, by Albert Sanchez-Graells
Emma’s reflections and questions condense what most of us have thought about and experienced when we first started considering the interactions between AI and public procurement. This is a quickly moving field and can feel daunting. There is a lot of hype, and it would be too easy to jump on the bandwagon of techno-optimism or, conversely, dismiss AI as ‘the new blockchain’ and a fad that will pass. It is important to ‘stick with the questions’ and develop that critical eye, but it can be difficult to decide where to look for accessible technical information and authoritative guidance.
Common to many challenges, it can be productive to take a step back and look at the bigger picture. AI is a data-intensive technology and has notable peculiarities but, in the broader scheme of things, AI is a new tool that needs to be understood before it is deployed. This can be done using adaptations of existing risk assessment and governance frameworks, and there is increasing guidance on how to make sure the socio-technical implications of AI deployment are understood and adequately mitigated. In the same way as every other tool it will take investment, training, and care to make the most of AI. A first step to consider is whether the ‘groundwork’ of ensuring the data and skills infrastructure required to buy and use AI are in place. Organisations can never start too early in getting their data and upskilling programmes in place.
In some fundamental respects, AI is ‘just another’ type of complex procurement of software, or the next iteration of tech procurement ‘after cloud’. This has the advantage of pointing us in the direction of challenges and solutions that have already been exposed to practical implementation and, in some cases, perhaps to interpretation, guidance, or even litigation. Not starting the process of thinking about buying AI and using it in procurement completely afresh can help us build on less shaky ground. This also has the advantage of facilitating some ‘leap frogging’ for those who had not directly engaged with cloud procurement, as there are clear pathways between complex procurement and AI procurement. We can then move to the more specific challenges posed by AI.
A Hitchhiker’s Guide?
In my recent book, I explore the two interactions between AI and procurement. I first look at the challenges of procuring AI. Most of them arise from the lack of regulation, experience, and even technical standards on AI (which are only starting to emerge). This makes AI procurement particularly challenging because it can feel to procurers that they need to ‘reinvent the wheel’. Although the new EU AI Act and the related model clauses for AI procurement provide a series of focus points to structure decision-making, the reality is that the burden is still on procurement officers to work out important aspects, such as the ways in which reliability, accuracy, or robustness will be assessed. There is also the additional challenge of operationalising those requirements in technical specifications, award criteria, and contractual clauses – and the broader complication of keeping an eye on potential improper users and failure points once the AI ‘hits the frontline’.
I also look at the use of AI in procurement. In my analysis, the true potential of AI lies with information-intensive tasks, such as the retrieval or direct comparison of data. However, AI can do little with the complexity and openness of many decision points in a procurement process. AI cannot replace discretion and judgement, and there are many open questions on how to adequately use AI to support human decision-making.
Common to both sets of interactions, AI poses challenges that may have no easy – or any – solutions (for example, around AI explainability), and these are issues that will require careful, ethical consideration.
It is clear that there is no simple ‘map’ or ‘how to’ guide when it comes to AI in procurement. However, by sharing reflections, learning, and experiences, we can jointly progress our understanding.
Explore more about AI and Public Procurement
For those interested in gaining a better understanding of AI and Public Procurement, consider attending our upcoming course Buying AI and Using AI in Procurement this December in Barcelona.
To read more about artificial intelligence, consider looking at our previous blogs on this topic by clicking here.
The views expressed in this blog are those of the authors and not necessarily those of EIPA.
