4299 Altus Insurance Whitepaper FINAL SPREAD - Flipbook - Page 30
Chapter 5: Regulator/Governance
Landscape (cont.)
Regulation as a disruptive influence
Regulation is no longer a slow-moving beast that
responds a昀琀er the fact. Today, it is a proactive and
increasingly disruptive influence in the market, not
afraid to make demands and to punish those who
don’t meet them. The defence against this remains
unchanged – e昀昀ective compliance – but the reporting
obligations make this a far more onerous task than it
has been.
Regulation is no longer a slow-moving
beast that responds a昀琀er the fact.
Today, it is a proactive and increasingly
disruptive influence in the market, not
afraid to make demands and to punish
those who don’t meet them.
AI ethics
At every turn, from the introduction of the GDPR/
UK Data Protection Act 2018 to the arrival of Fair
Value rules, insurers have turned to technology, and
understandably so. AI and machine learning are the
perfect tools to identify, gather, clean and report
the necessary data, and as those data and reporting
demands increase, the expectation is that technology
will adapt to respond.
Unfortunately, while technology might secure some
crucial breathing space, it only really addresses the
problem of scale. As more data is collated and more
AI employed to manage it, new regulatory risks are
created. The use of technology to meet the regulatory
requirements is, ironically, creating even greater
regulatory scrutiny.
The application of AI to a growing list of activities
has raised concerns around the ethics of its use.
The fairness of decisions made by AI is a crucial
consideration, as baked-in biases could result in
adverse outcomes for customers and failure to comply
with FCA rules.
“Do no damage to humans.”
Overwhelming sentiment from all AI
Regulatory guidance to date
Accountability
& governance
Safety,
security &
robustness
Contestability
& redress
Principles
Fairness
Appropriate
transparency &
explainability
Underwriter considerations
• The application of AI models to support
underwriting is corporately governed
and reviewed
• Any use of AI in underwriting appetite
or decisions must be documented and
explainable
• The use of AI must be able to be proved
it doesn’t create anomalies or bias,
given consistent underwriting inputs
• The use and application must be able to
be defended and to be adjusted
Figure 5.2: Ethics of Arti昀椀cial Intelligence
30
