How AI Solutions And Strong Leadership Could Have Avoided The UK’s Worst-Ever Miscarriage Of Justice

Fundamental digital modifications are upon us to combine clever options in enterprise.


The 14-year-long prosecution of a whole bunch of workers within the United Kingdom, dubbed the Post Office Horizon scandal, shook a world IT and AI neighborhood of which I’m an element. This was as a result of this miscarriage of justice was based on data from an IT laptop system baptized Horizon. I’ve spent years creating what I coin “clever anomaly detection methods” each within the United States and in France the place my staff and I develop AI instruments to suggest options and safeguards to real-world challenges in enterprise. I admit to being baffled on the Post Office’s belligerence in wrongly accusing 736 workers of misreporting and theft as a substitute of questioning the chance that Horizon, designed by Fujitsu, could also be malfunctioning. This scandal made me mirror extra about not solely the design and deployment of IT methods and AI options, but additionally doable methods to handle the chance of utilizing these technological options and find out how to construct accountability into these options.

With hindsight, many managerial and management actions may have been applied to forestall a scandal which has destroyed a whole bunch of lives. To start with, we must always have a look at it by means of the chance administration perspective. The Post Office ought to have set a transparent tone on the high concerning the transparency and integrity of the IT methods put into place. It may have performed an intensive investigation of any potential defects within the Horizon system earlier than signing the contract with Fujitsu. Furthermore, it will have been properly suggested to make a strong danger administration and monitoring plan of the system after its implementation.

Moreover, the Post Office ought to have taken the queries by the whistleblowing sub-postmaster, Alan Bates, extra critically when he requested for entry to the information to attempt to perceive the discrepancies in his department. In 2000, Bates reported the issues linked to Horizon to the Post Office administration staff. Unfortunately, his studies have been dismissed and his contract with the Post Office was terminated three years later. In its Bates & Others vs Post Office Limited judgment, the High Court of Justice acknowledged that the Post Office’s method was “a easy institutional obstinacy or refusal to contemplate any doable alternate options to their view of Horizon”. It went on to conclude: “It quantities to the twenty first century equal of sustaining that the earth is flat.”

Detecting And Flagging Anomalies With AI

We imagine the AI options now we have developed may keep away from these obsolescent attitudes. For years, now we have been engaged on an anomaly detection system designed for inside audit, danger administration, and compliance functions. When you set accounting information into the detection system, it could actually assign anomaly scores to every monetary transaction. It tells you why sure transactions or patterns of transactions are anomalous-based on the metadata and the accounting construction of transactions. In this case, our detection system ought to be capable of detect the modifications in transaction patterns after implementing the Horizon system and flag lots of the incorrect information generated by Horizon as extremely anomalous. Furthermore, our algorithm also can generate explanations regarding how the anomaly scores have been assigned based mostly on the algorithm’s logic. Incidentally, our graph-based machine studying algorithms can be used to counter anti-money laundering, in addition to e-mail communication monitoring, and fraud detection functions.

But nonetheless, we would want the Post Office administration to take the purple flags of the Horizon system critically and examine accordingly. After all, this miscarriage of justice isn’t solely a few flawed IT system, it is usually about how the Post Office handled it.

Fundamental Changes For More Robust Cloud-based POS Systems
We are additionally satisfied there’s a helpful management lesson for Fujitsu. We imagine the multinational firm mustn’t solely be more practical in decreasing product defects but additionally have a look at how its purchasers are utilizing the output of its methods. Horizon is a domestically based mostly point-of-sale (POS) system that information transactions, however the Post Office additionally used the information output to prosecute its workers. More consideration ought to have been paid to the information output at that time. Perhaps, Fujitsu mustn’t have handed over information packs to the UK Post Office as courtroom proof.

Instead, there seems to be a travesty of justice surrounding Fujitsu’s involvement within the Horizon scandal. After a decade of investigation, award-winning reporter Nick Wallis claims in his e-book “The Great Post Office Scandal” that the Japanese know-how company was no bit-part participant, however “on the coronary heart of the scandal” as a result of defective retail and accounting software program it created and put in. Yet, Fujitsu has not solely refused to pay any compensation cash to the victims, it has simply signed two large UK authorities IT contracts value over $520 million. This has generated robust criticism of the British authorities – even from inside its personal ranks the place one Conservative peer stated: “The authorities needs to behave as if it’s ‘enterprise as normal’ with Fujitsu. What is the motivation on Fujitsu to contribute to the huge compensation value about to fall on the taxpayer?”

Eyes On The Cloud
More typically, elementary digital modifications are upon us to combine clever options in enterprise. More and extra main corporations in retail are adopting cloud-based POS methods built-in with strong cloud-based Enterprise Resource Planning (ERP) methods. They are extra user-friendly and companies can embed highly effective AI into their ERP methods. These, for instance, can embed algorithms into ERP constructions to determine issues within the monetary flows. Thus, it ought to substitute conventional money register methods like Horizon. Many companies are additionally contemplating integrating blockchains into their ERP methods to enhance the integrity of transaction information. However, including blockchain know-how requires extra work in constructing the infrastructure and collaborating with enterprise companions.
All these technological explorations are on the coronary heart of our analysis mission on the HI! Paris Center on Data Analytics and Artificial Intelligence. This includes creating machine learning-based instruments to enhance our understanding of accounting information. It is analysis that seeks clever options to real-life challenges like those we noticed within the UK Post Office affair. In this manner, we hope that, by no means once more, will a scarcity of rigorous monitoring of the IT system result in such travesties of justice by which a whole bunch of harmless staff are wrongfully prosecuted.

Aluna Wang is HEC Assistant Professor of Accounting and Management Control. She can be chairholder at HI! Paris Center on Data Analytics and Artificial Intelligence.
Daniel Brown is Chief Editor at HEC Paris.

Recommended For You