For three years, the Basel Institute’s Prevention, Research and Innovation team worked with 24 other research partners from across Europe on the ambitious Horizon Europe FALCON (Fight Against Large-scale Corruption and Organised Crime Networks) project.

Our consortium brought together expertise in social science and research, technology development and law enforcement. We worked closely together to innovate the fight against corruption and organised crime by designing data-driven risk indicators and AI-enhanced solutions.

This massive research project generated many useful “deliverables” – corruption intelligence, risk indicators, practical tools, policy guidance – that anti-corruption and counter-crime practitioners can directly apply in their work.

Perhaps more importantly, the research and interactions have helped us all understand corruption better – and provided the evidence we need to innovate anti-corruption methods.

This is bringing real benefits to our technical assistance to governments and organisations seeking to find smarter ways to prevent and combat corrupt and criminal networks, in Europe and beyond. It’s an example of why international, multidisciplinary anti-corruption research projects are worth the investment even in times of tightened belts and budget cuts.

As the project winds down, here are four lessons we’d like to share with others. We also offer our warm thanks to the consortium leaders and partners, and to the EU institutions and individuals who continue to promote deep research into corruption and crime.

Lesson 1: Understanding how corruption evolves helps anti-corruption practitioners keep up

FALCON provided clear examples of how – and how quickly – corruption and crime adapt and evolve.

Traditional understandings of corruption focus on a fairly static list of behaviours like bribery or embezzlement. In fact, FALCON research demonstrated how corruption occurs in complex systems, playing out through resilient networks of actors, institutions and incentives that evolve within equally complex political, economic and social contexts.

Each new enforcement measure, detection tool or regulation changes the behaviour of those trying to circumvent it. When challenged, criminal actors adapt, prompting authorities to respond with further reforms.

It’s a never-ending dance of push and shove that we observed in a range of different settings. For example, in our research for the FALCON project, we have seen that criminal actors:

• adapt their corruption and criminal strategies in response to strengthened enforcement (see our articles on the evolution of corruption and on corruption and crime at the Kapitan Andreevo border);
• relocate their illicit activities elsewhere (see our Working Paper 58 on corruption as a facilitator of drug trafficking in the port of Rotterdam); and
• apply counter strategies, such as undermining reform efforts and dismantling or weakening anti-corruption institutions and policies; or even
• weaponise the anti-corruption apparatus against rivals (look out for our forthcoming report on backlashes to anti-corruption reforms).

How to keep up?

Our approach, which we refined during the FALCON project, is to always base anti-corruption and counter-crime interventions on a strong understanding of underlying criminal dynamics.

Map the stakeholders and their relationships, interests and incentives. Identify risks and problems that could arise.

From there, legal reforms, grassroots initiatives, technology and everything else can function best as a flexible anti-corruption toolbox. We assemble and revisit tools as the context shifts and our knowledge advances.

Lesson 2. Use AI, but understand its limitations

A key point often forgotten amongst the technology optimism is to treat tech tools, including the ones built as part of the FALCON project, as inputs to expert human judgement, not as substitutes for it.

A risk indicator or dashboard score can flag suspicious activities that require a closer look, but it can’t tell you why an anomalous pattern exists or what to do about it. We must check an AI model’s output against contextual knowledge and other evidence, and keep in mind what it was and wasn’t calibrated to detect.

The same caution applies to AI. AI tools can be useful for analysing huge quantities of data and detecting patterns that point to corruption. But an AI could just as easily facilitate corruption as help you to tackle it, so make sure human judgement stays in the driving seat.

Lesson 3. Reach beyond the bubble to identify synergies or unintended consequences

It can be tempting for anti-corruption researchers to work only within the anti-corruption community, and may seem more efficient.

But the truly multidisciplinary FALCON research consortium showed how closely crimes and corruption are interlinked. Look at them separately and you miss the true picture. Intervene to prevent one problem, and you might have unintended impacts on another.

An example from our research at the border between Türkiye and Bulgaria:

• Beefing up border control capacity against trafficking can result in criminals relying even more on corruption to avoid detection.
• Conversely, stronger enforcement against corruption can lead to traffickers using more sophisticated criminal strategies to get illicit goods safely across the border.

Before designing any intervention, we should assess how increased pressure in one policy area may affect others, including through the displacing of risks.

Anti-corruption programmes are more effective when we ask these questions from the outset and coordinate more meaningfully with those working on related issues like organised crime and security.

Lesson 4. Use strategic foresight – systematically anticipate new risks and changed contexts

Our work on foresight and anticipatory governance has enhanced how we approach political economy assessments, including in projects unrelated to FALCON.

We have long found it important to consider what could shift the terrain – a new law, technology, conflict or other – and what this would mean for a planned anti-corruption intervention. Now we have found tools and systems to do this systematically and periodically.

The approach helps our partners to revisit anti-corruption frameworks, track emerging risks and anticipate unintended consequences. That’s key to mitigating them and adapting in good time.

It also tangibly improves monitoring and evaluation frameworks, as we are not only reporting on what happened in the past, but adapting to what is up ahead.

What’s next?

The FALCON research project provided not only the practical guidance and reports it promised, but also a far more detailed understanding of how corruption and crime interlink and smarter tools to act on them both.

As the project closes, we are grateful for the partnership, energised by what we’ve learned and ready to expand and apply them in our technical assistance to governments and development partners worldwide.

The FALCON project was funded under the European Union’s Horizon Europe Framework Program Grant Agreement ID 101121281. The Basel Institute’s involvement in the project was funded by the Swiss State Secretariat for Education, Research and Innovation (SERI). Views and opinions expressed are those of the author(s) only and do not necessarily reflect those of the European Union, the European Research Executive Agency or SERI.