Artificial intelligence – will it help or hinder the world of fraud?
In recent weeks, reports have emerged of prominent companies and organisations using artificial intelligence (AI) to crack down on financial crimes. The Serious Fraud Office (SFO) reportedly hired an “AI lawyer” to assist in analysing complex documents in its recent case against Rolls-Royce, while HSBC is allegedly investing money in “bringing in robots” to help spot money laundering and fraud in the near future, The Financial Times reports.
The move towards normalising the use of AI in complex investigations is intended to reduce costs and to tackle financial crime quickly and more efficiently. According to the Financial Conduct Authority (FCA), banks, for example, spend £5 billion a year on tackling financial crime in the UK alone – a figure which HSBC seems to think can be significantly reduced by taking advantage of the latest technology.
According to reports, Europe’s biggest bank is hoping to strike-up a deal with prominent AI software firm Quantexa. The move comes after years of hefty fines imposed on banks over failures adequately to prevent illegal funds from flowing through their accounts – an issue which is likely to grow in prominence over the coming years as criminals themselves seek out the latest ways to use modern technology to their advantage.
Earlier this year, HSBC successfully froze an account believed to be connected to an alleged $500 million fraud carried out by the son of the former president of Angola – a move the bank claims is indicative that its recent efforts to clamp down on compliance have so far been successful. Nevertheless, it is evident that in this day and age, further innovation is needed to combat the increasingly sophisticated nature of fraud.
AI experts such as Quantexa believe that the only way to counter emerging threats is by “harnessing technology and data” in order to reach an ideal future whereby banks can “understand and see criminal behaviour in as near real-time as possible,” and HSBC is pumping billions of pounds into its compliance controls in a bid to ensure that its own systems edge ever-closer to this idyllic reality.
These efforts may indeed help to streamline fraud detection and prevention for banks. However, when it comes to more in-depth investigations, such as the SFO’s probe into Rolls-Royce, things get a little more complicated and human assessment remains vital.
The SFO estimates that using so-called “AI lawyers” to analyse documents as part of fraud investigations is some 80 per cent cheaper than using outside counsel to identify legally privileged material within those documents. However, OpenText, a Canadian company that develops “AI lawyers” and similar technology, says that the legally privileged material such technology is capable of “flagging” still nee