AI robot

Artificial intelligence can identify risks in public contracting – court

Algorithms will detect “contracts favouring same businesses, excessive costs and conflicts of interest”

At a moment in the country’s political reality where ‘scandal’ at local and national level is never far from the headlines, an ambitious new project has been launched in Portugal to use artificial intelligence to identify risks in public contracting.

The project was presented today, in its initial phase, at a conference in Lisbon, bringing together the Court of Auditors, the OECD and the European Commission (which is funding the plan). 

Said  Helena Abreu Lopes, one of the judges responsible for the system’s implementation: “This is very complex, very ambitious. We have to be aware that the project won’t manage everything in two years. It is a work in progress that will have to be deepened”, but it aims to “make it possible to identify patterns of behaviour, the awarding of the same products to the same companies, and even employees in situations of conflicts of interest”.

Said the judge, machine-learning algorithms will be able to run through the 2,000 to 4,000 contracts the court has to analyse every year, flagging any risks of fraud or misuse of public resources.

Keeping the discourse as “politically neutral” as possible, José Tavares, chairman of the Court of Auditors, said the “expected long-term effect of the project (impact) is the improvement in the identification of risks/ unusual transactions, the improvement in the early detection of potential irregularities and real-time monitoring”.

Algorithms, after all, are not influenced by outside pressures. They cannot be wined and dined, offered luxury holidays or any other favours. They simply work ‘as programmed’, which José Tavares believes will enable “greater capacity for control and more efficient allocation of resources by the Court which has to oversee the contracts of around 6,500 public entities.

As Lusa explains, 25% to 30% of public spending “is the result of public contracting”. The State news agency hasn’t given a figure for ‘direct contracts’ (those that do not go out to tender), and José Tavares took care here not to wade into deep water.

When questioned about the impact the project may have in terms of detecting risks or irregularities in the scope of direct negotiation contracts, José Tavares stressed that direct negotiation “is sometimes spoken of in a way that is not the most correct, bearing in mind that in many situations it is preceded by market consultation.”

He also “dismissed the idea that direct negotiation can be identified with corruption” (…) This phenomenon of corruption can take place either in direct negotiation contracts or in competitive procedures such as public tender,” he told Lusa.

Thus all the more reason perhaps for bringing a possé of algorithms into the mix.

According to Jornal de Notícias, “with artificial intelligence, the Court of Auditors will be rapidly able to evaluate between the (thousands of) contracts it receives every year for prior inspection, price deviations, excessive pricing, compliance with deadlines and even the degree of compliance with environmental data”.

All this, without neglecting “privacy, ethics, confidentiality, data protection and cybersecurity”, guaranteed José Tavares.

[email protected]