Serious Fraud Office hires AI lawyer: a Rolls Royce solution?

Serious Fraud Office hires AI lawyer: a Rolls Royce solution?

AI that can work 2,000 times faster than a human lawyer has been deployed to help tackle fraud and corruption in the UK.

The Serious Fraud Office (SFO), the prosecuting authority that tackles top-level fraud, bribery, and corruption cases, has deployed an artificial intelligence (AI) ‘robo-lawyer’ to speed up document analysis and improve accuracy.

The SFO is part of the UK’s criminal justice system, covering England, Wales, and Northern Ireland, but not Scotland, the Isle of Man, or the Channel Islands.

The organisation has the unenviable task of processing a huge number of documents every day, and needed to upgrade its document analysis capability. The new system, Axcelerate, will be provided by software company OpenText, and will enable the SFO to investigate cases more quickly, reduce costs, and work at a lower error rate than its human legal teams are able to do.

The SFO recently piloted a software robot to process more than half a million documents a day. It was used to scan privileged content in the SFO’s recent Rolls-Royce case – the biggest investigation the organisation has ever had to undertake.

According to the SFO, the AI system was able to work 2,000 times faster than its human legal counterparts.

Rolls-Royce had to apologise after the SFO found conspiracy to corrupt, or failure to prevent bribery, by the car manufacturer in China, India, and other markets. The company was made to pay £671 million to settle corruption cases with UK, US, and other authorities.

The SFO now hopes that the Axcelerate system will help case teams to target their time and efforts more efficiently, by pursuing other aspects of investigative and prosecutorial work while the AI shoulders the document discovery workload.

Previously, only independent barristers had the experience to comb through thousands of complex documents to identify evidence for SFO investigators, prior to them sifting through the documents themselves.

The new technology will help the SFO to recognise patterns, group data by subject, organise timelines, and de-dupe information. According to the SFO, it will eventually be able to sift for relevancy – meaning that irrelevant documents will be removed from investigations quickly, potentially saving huge amounts of time.

“AI technology will help us to work smarter and faster, and to more effectively investigate and prosecute economic crime,” said SFO chief technology officer, Ben Denison.

“Using innovative technology like this is no longer optional – it is essential given the volume of material we are dealing with, and will help ensure that we can continue to meet our disclosure obligations and deliver justice sooner, at significantly lower cost.

“The amount of data handled by our digital forensics team has quadrupled in the last year and that trend is continuing upwards as company data grows ever larger,” he added.

Internet of Business says

This observation that data volumes are booming in business is the key to solving investigators’ problems: increasingly, it is getting harder for human beings to deal with the vast amounts of data that exist in organisations’ data centres, as well as in their filing cabinets.

AI’s ability to recognise patterns in data or behaviour – at high speed – promises to improve countless processes, and help a range of organisations to spot illegal behaviour, diagnose or predict diseases, identify new business opportunities, and kickstart a range of other opportunities.

The announcement comes in the wake of a number of other applications of AI to help organisations identify and investigate fraud and other illegal activities (see sample links below).

But in routine work, some organisations should consider another possible impact: on human skills. In the legal sector, for example, many newly qualified lawyers learn the ropes of their profession by doing bread and butter work, such as document discovery and routine investigations.

Wholesale, tactical automation of these processes might help experienced lawyers to serve their clients even better – and conceivably at lower cost or higher profit – but there is a risk it might chip away at the first rung of the career ladder, for those who need case work to match their years of theoretical training.

Read more: HSBC employs AI fraud detection in $2.3bn tech spree

Read more: NatWest prevented £7m of corporate fraud using machine learning

Read more: Anti-doping agency to use AI to catch sports drug cheats 

Read more: Fintech firm unveils AI-powered mortgage broker

Read more: Apple hires Google AI chief to head machine learning