Technological change - automation and digitisation, big data, artificial intelligence and cyber risk

Key Developments During 2018

Technological Change

  • Technology is driving change in the economy with a recent survey indicating that data scientists will be the most important workers in the search for future growth, with emerging technology specialists such as AI professionals close behind. These roles will make use of automation and digitisation, big data and artificial intelligence and may increase challenges such as modelling risks for which there is no existing data or the liability or impact is unclear (e.g. autonomous vehicles, genetic modification).
  • The FRC’s Lab has been investigating the impact of new technologies through its Digital Futures project. It has issued two reports “XBRL: Deep-dive Digital future of corporate reporting” and “Blockchain and the future of corporate reporting: How does it measure up?”. A final report on artificial intelligence will be released in early 2019.

Operational Challenges

  • u  Financial services firms are increasing their reliance and dependence on technology and may face operational challenges in adopting new technologies whilst maintaining legacy systems, adapting to complexity of Brexit or facing cyber incidents. High profile operational incidents include RBS, TSB and cyber attacks on NHS, Bangladesh Central Bank, Equifax, Yahoo, and Sony. In July 2018, a joint discussion paper was issued by the PRA, the FCA and the Bank of England “Building the UK financial sector’s operational resilience”. The purpose of this DP is to share the supervisory authorities’ thinking regarding operational resilience and obtain feedback. 

Actuarial Considerations

  • KPMG report that “AI systems may be seen as a ‘black box’, making important decisions when few people outside of analytics teams, data science labs and technology centers can fully understand how.” The appropriate use of models and understanding of ‘black box’ models including ethical challenges continue to be debated by the industry and regulators. The PRA has issued a supervisory statement “Algorithmic trading” in June 2018 regarding banking firms’ governance and risk management of algorithmic trading. Increased automation and use of artificial intelligence within actuarial work may change the nature of actuarial work with implications for regulation and education.
  • The ethical questions raised by big data research have increased as the size and complexity of available datasets has grown and big data research is woven into our daily lives (e.g. mining digital medical records for scientific and economic insights, mapping relationships via social media, capturing individuals’ speech and action via sensors, shaping security policy via ‘predictive policing’, etc). These techniques are also increasing within the insurance and savings industries. Therefore, actuaries should be aware of data privacy and security issues (e.g. GDPR obligations) and ethical considerations (e.g. using geocoding to infer health characteristics) as well as any accuracy or reliability limitations (e.g. third party or scraped data). In May 2018, the Royal Statistical Society and IFoA announced plans to collaborate on the exploration of the practical and ethical implications in the rapidly developing field of data science.
  • In its review of pricing in household insurance the FCA found that there was a risk of discrimination against consumers through using rating factors in pricing based on data relating to or derived from protected characteristics and of differential pricing leading to some identifiable groups paying significantly higher prices than others with similar risk and cost-to-serve characteristics. The FCA has decided that a package of measures is necessary on insurance pricing (see Financial Security hotspot) and carried out a market study into how GI firms charge their customers for home and motor insurance which closed on 3 December 2018.

Summary of 2017 Discussion and Actuarial Implications

  • Technological advances and developments in AI are changing the nature of insured risks and financial products, as well as the tools used to model and manage them. Recent reports and surveys identify the emergence of big data and analytics as a key trend and area of investment, especially in the insurance sector.
  • These developments may result in disruptions to current business models, may invalidate the judgements, models and assumptions currently used by actuaries or make it more difficult for actuaries to interpret data and communicate insight. Actuaries need to make sure they are aware of how the new technologies are impacting the underlying risk and adapt their methods, models and assumptions using the latest tools – where appropriate.
  • For example, the increased use of telematics in motor shows how technology can inform and influence policyholder’s behaviour and give flexibility in pricing, product design and added services. Where changed driving behaviour changes the frequency and severity of claims, past data may no longer be appropriate for pricing or reserving. Enabled by the detailed telemetry data actuaries can, using increased computing power and more flexible analysis tools, investigate trends or risk factors which it was not possible to even know about before.
  • Some data sources or practices may pose ethical questions. For example, using data relating to health or lifestyle or using data for pricing in relation to propensity to switch that may unfairly disadvantage some groups. Actuaries can play a role in decisions around the suitability of data use on technical and ethical grounds and support its use in the public interest.
  • Increasing digitisation and connectivity creates new cyber risks, such as data loss and cyber crime. These may impact actuaries themselves, e.g. data corruption leading to advice based on incomplete data or the industry e.g. insurers need to develop products and methods for pricing and reserving for cyber exposures whilst avoiding unintended coverage.
  • As well as these risks there are opportunities for new products and new business models where actuaries’ skills and experience can help users take decisions and understand uncertainties.
  • There are sources of information and research available within the profession such as the MAID working party and many external sources – some of which are listed in the Risk Perspective reference section.

Further Reading

Previous Hotspot - Pension Scheme Management / Next Hotspot - Terrorism and Cyber Crime