Interest in AI in the Middle East has surged in recent years, with several governments announcing national AI strategies and implementing initiatives to drive its adoption. Norton Rose Fulbright listed some as including:
- In 2017, the UAE appointed a minister for AI
- In 2018, the UAE published its national strategy for AI 2031. Following this, the UAE has issued various AI guidance notes
- In 2019, it launched the world's first AI university.
Similarly, Saudi Arabia formed the Saudi Data & AI Authority in 2019 to serve its goals under Vision 2030 and has plans to invest $20bn in AI by 2030. Initiatives in other countries have also followed, with countries such as Qatar and Egypt also unveiling ambitious AI strategies and e-government programmes.
"The expected financial impact of AI in the Middle East is significant, with the region estimated to accrue $320bn in value added by AI by 2030," according to Norton Rose Fulbright partner Shabnam Karim.
The integration of AI has become prevalent in over 50% of
insurance customer interactions in Saudi Arabia, marking a
significant shift towards
efficiency and innovation in the industry.
It was reported that the country recorded more than 80m transactions within the insurance sector through AI.
Business integration
AI is being employed for a wide range of applications within the MENA insurance sector from underwriting and risk assessment, to claims processing and customer service.
Ms Karim noted that Saudi Arabia's insurance industry saw over 50% of customer service interactions in 2023 conducted through AI, according to data from the General Organisation for Social Insurance.
In Algeria, insurance company Macir Vie has developed an AI-based chatbot called Hayat that provides tailored advice and information sharing to customers, while in the UAE, insurance company I-Insured has launched an AI-based SafeDriver app that offers policyholders incentives and premiums based on their safety on the road, Ms Karim said.
“As part of the customer service line, AI is also being trained to detect sentiment and convey empathy and to help agents formulate personalised offers to insurance customers,” she said.
In underwriting, AI has ushered in a new era of streamlined data-driven decision making.
“Machine learning is being used by insurers to pore through vast datasets to assist in identifying risk factors and for pricing. For instance, in the UAE, Union Insurance Company is using natural language processing technology to analyse data from documents uploaded by customers and issue motor policies in less than one minute,” she said.
AI is being implemented by insurers in the region to prevent fraud and reduce costs.
“Industry estimates suggest that abuse of health services in the UAE costs approximately 10-15% of total collected premiums. Predictive analytics can review historical data and identify patterns that are indicative of potentially suspicious activity to alert insurers to fraud in real time.”
“AI capabilities may also be employed to streamline and automate claims processes, particularly for more repetitive or high-volume insurance claims,” Ms Karim said.
Challenges
Despite its potential benefits and use cases, the adoption of AI by insurers in the MENA region presents certain challenges that may affect its uptake if not addressed effectively.
Norton Rose Fulbright senior associate Shiv Daddar said various jurisdictions in the MENA region have enacted robust, modern data privacy regimes, including Saudi Arabia, the UAE, DIFC and ADGM. These place stringent rules on the use of personal data within AI systems, including the requirements to process personal data fairly, lawfully and transparently, according to spokespeople from Norton Rose Fulbright.
“In practice, this requires insurers to provide meaningful information to individuals about the logic and consequences of AI decisions. Insurers will need to prioritise the ‘explainability' of AI solutions, particularly as these systems automate decisions that would have previously been taken by humans,” Mr Daddar said.
“The adoption of AI also poses ethical questions around unfair algorithmic bias. A notable risk is that the algorithms may unintentionally replicate historical biases present in the data they are trained on, which could lead to outcomes that discriminate against specific group,” he said.
Left to its own devices, there is risk in an AI model having no safety, legality or morality constraints.
“A form of human oversight is often therefore required to compensate for these flaws and to monitor the accuracy and good functioning of the model. Steps will need to be taken to ensure that the training data is of high quality. These may include removing examples from training datasets that the insurer considers may lead to discrimination, testing the output and then modifying the model after the initial training by applying the learnings of these outputs,” he said.
This is particularly important for insurers given the real-life consequences that decisions based on this data have on insureds and related parties.
Other challenges include issues around ownership of outputs of AI systems and allocations of risks and liabilities when things go wrong, particularly where all or part of an AI system incorporate third-party solutions, Mr Daddar said.
When asked what measures can be taken to mitigate the impact of AI on employment, Norton Rose Fulbright associate Simon Lamb said, “Whilst the deployment of AI can lead to increased productivity and efficiency, it also carries the potential to displace human labour and widen industry skill gaps.”
“There is an imperative for companies and governments to create programmes to facilitate job transitions and retraining to ensure that the workforce is equipped for an AI-driven future. To bridge this skill gap, the Saudi government has established the Prince Mohammed bin Salman College of Cyber Security, while the UAE has launched the world's first university for AI,” Mr Lamb said.
These measures are directed at fostering strong knowledge-based economies with talent in AI-related fields.
Driving the next wave
“It is expected that insurance companies will increasingly leverage AI to offer individualised policies to customers based on their personal behaviours, preferences and needs,” Norton Rose Fulbright partner Marcus Evans said.
Mr Evans noted as an example, in the health insurance sector, AI algorithms can analyse vast amounts of medical data and patterns to predict the likelihood of medical conditions and tailor insurance products accordingly.
Another example he noted was that the internet of things may be used to replace traditional, manual methods of first notice of loss.
“Natural language processing in Arabic is rapidly developing. For example, NOOR- developed by the Abu Dhabi Technology Innovation Institute- can extract information from texts and facilitate translations of documents. Given the often-significant costs and time associated with translating and assessing coverage under insurance policies, this may result in significant efficiencies for insurers in the region,” Mr Evans said.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.