New technologies - Machine learning and Artificial Intelligence

For the moment, the legal framework covering insurance industry activity does not regulate the use of Artificial Intelligence (“AI”) or machine learning (“ML”).

For the moment, French legislation does not regulate the use of AI or machine Learning. Parliamentary reports have been published but have not yet been followed by any action.

Several points have been debated, such as the standardisation of customer identification procedures or those for signing contracts online. In addition, while the use of new technologies such as AI, blockchain or Application Programming Interfaces, meaning the connecting nodes of the digital society, thus a foundational component of its digital transformation (“APIs”) is increasingly widespread, there is still no official guidance from the authorities and neither from the legislator on the correct use of such tools. IVASS has recently launched a consultation on the use of machine learning algorithms in insurance sector.

Currently there is no specific insurance regulation relating to the use of machine learning and AI nor any sectorial code of conduct on this matter.

Consequently, when deploying these technologies in the development and distribution of insurance products, entities should be mindful of their duty to act in customers’ best interests and be in compliance with the information duties, in accordance with the Consumer Act (Law no. 24/96, of 31 July1996), namely fairness and transparency principles.

The European Insurance and Occupational Pensions Authority (“EIOPA”) has published a report on digital ethics setting out AI governance principles for an ethical and trustworthy AI in the European insurance sector. Thus, being mostly principles-based, this document sets forth the key elements for implementing the impact assessments within AI- driven tools in Insurance.

Additionally, the EU’s proposal on AI Regulation which is expected to come into force during 2022, should be taken into account. This sets the grounds for a risk-based approach towards AI applications, which must, consider the high risks, automation and optimised pricing approaches in the insurance sector.

On 25 July 2019, the Dutch supervisory authorities and the Dutch Central Bank, in Dutch “Autoriteit Financiële Markten” (“AFM”) and De Nederlandsche Bank (“DNB”), have jointly published an exploratory study on “Artificial intelligence in the insurance sector”. The study contains 10 points of consideration when applying AI in the insurance sector.
Furthermore, DNB published in April 2020 its “Theme research Insurtech”, in which it outlines the results of an inventory of Insurtech within small and medium-sized companies. DNB considers it important that insurers are aware of the potential opportunities and risks of Insurtech.
On 26 January 2022, DNB announced that due to the quick technological developments it will continue to follow AI developments, with respect to the insurance sector. The DNB project “AI and rationale” will be continued in 2022, and DNB will be actively involved in international meetings about AI, for example, via EIOPA.

There is currently no specific legislation, guidance from the authority, or code of conduct regulating the use of ML or AI.

At present, there is no specific insurance regulation relating to the use of machine learning and AI nor any specific code of conduct.

The UK regulator has recognised the benefits of digital innovation for consumers, including those offered by machine learning and AI, and has stressed the need for transparent, fair, secure and ethical use of such technologies. It has established an AI Public-Private Form which seeks to support the safe adoption of these technologies within financial services and to consider whether principles, guidance, regulation and or industry good practice could support this.

When using such technologies in the development and distribution of insurance products firms should be mindful of their duty to act in customers’ best interests and the rules and guidance around ensuring fair value. The introduction of the new Consumer Duty in will also be pertinent to firms’ use of machine learning and AI. The UK regulator has also stressed that firms using these technologies are expected to fully understand such technology and to have robust governance in place.

The UK regulator is supportive of innovation within financial services and its Innovation Hub offers various services to insurance market participants that are seeking to use technology such as machine learning and AI in an innovative manner. This includes access to a ‘regulatory sandbox’ in which new propositions may be tested in a live environment with relaxations to certain rules for a limited time.

It is also worth noting that the UK government published its National AI Strategy in September 2021. This sets out the UK’s intent to become a world leader in AI based around the three pillars of investment, ensuring AI benefits all sectors and regions, and governance. This has a broader reach than the EU’s AI Regulation but is less precise in its proposals.

The Spanish insurers’ Association (“UNESPA”) has created a Code named, “Principles for the ethical use of AI in the insurance sector”, allows entities voluntarily adhere and comply with the principles contained in the Code.

Moreover, for Artificial Intelligence, the following developments may apply:

• Spain has proposed to the Commission to become a testing platform for the AI Regulation. Through a pilot test, the European standard would begin to be applied with certain companies or industry participants before it becomes mandatory, as a pilot test. Currently, the plan is in the negotiation phase with the Commission's Joint Research Centre (“JRC”);
• On 14 July 2021, the Digital Rights Charter was issued by the Spanish Government, which is not regulatory in nature, but it proposes a reference framework and lays the groundwork for future regulations;
• During July 2022, the new Digital Agenda called “Digital Spain 2026” was issued, which includes a set of measures including the objective of promoting the transition to a data economy, guaranteeing security and privacy, and taking advantage of the opportunities offered by Artificial Intelligence;
• Also, in July 2021, the Government launched the Observatory on the Social Impact of Algorithms (“Obisal”);
• Consequently, the following measures are foreseen to be developed: (i) National Artificial Intelligence Strategy; (ii) Data Office and Chief Data Officer (“CDO”); (iii) Artificial Intelligence Advisory Council; and (iv) Cloud Strategy: European shared data spaces.