AI will be used by CSPs as a superhuman shield against fraud
There’s been a lot of hype around machine learning (ML), and artificial intelligence (AI) in recent years, and in 2019, we believe ML and AI will be used as tools to help fight telecoms fraud, writes Daniel Kurgan, the chief executive of BICS. It’s a damaging and persistent issue, which costs the international wholesale carrier industry US$17 billion annually. The automation and problem-solving which ML and AI can bring will help communications service providers (CSPs) stay ahead of fraudsters, delivering above and beyond what’s been humanly possible in the past.
The technology will be adopted by CSPs to identify more fraudulent attacks on their network, automating laborious manual processes, and analysing traffic in near-real time. AI-powered fraud-detection solutions have another superhuman quality: 24/7 working hours. Evenings and weekends are currently primetime for those committing telecoms fraud, as fewer employees are working and available to monitor networks.
Real-time – and all of the time -detection of problems will reduce the time it takes for CSPs to address and resolve problems, and therefore push down the value of each attack to criminals – and the financial loss to CSP. So, while we may well see the number of attacks rising in 2019, we predict – and very much hope – that the value of these attacks will fall.
Of course, any adoption of AI and ML in fraud-prevention solutions must be accompanied by human knowledge and expertise. We see a far greater number of fraud attacks targeting international, rather than domestic, network traffic. So, while CSPs might know their subscribers and their usual behaviour pattern when looking at domestic traffic, malicious activity will likely be more difficult to identify on international traffic – where it matters most. AI and ML has to learn from somewhere, so we think solutions of this type will only develop during 2019, when combined with human input and knowledge of the fraud threat landscape.
To support any investments made in ML and AI this year, the industry must remember: there’s strength in numbers. We need more parties to buck a trend, which until recently has involved most telecoms players staying relatively silent on the issue of fraud. Granted, few CSPs will be happy shouting from the rooftops about criminal activity – and financial losses – on their network. However, collaborating, sharing insights and resources, and contributing to dedicated initiatives to build knowledge and hasten detection, will ultimately reduce telecoms fraud.