Next time you contact your insurer there is a good chance that you will be dealing with Artificial Intelligence (AI). Insurers are starting to use intelligent computers to give advice and speed up the process of buying insurance and paying claims. Virtual underwriters and risk advisors are no longer confined to insurers’ research and development departments. But with the benefits also come new risks.
The age of machine learning
Artificial Intelligence (AI), also referred to as machine learning, is essentially software that is able to think and learn like a human. Today, basic forms of the technology are able to perform specific tasks – like checking an insurance claim for fraud – but future generations of AI will be capable of solving complex problems and making decisions, in much the way humans do today.
As such AI could double the annual economic growth rate in 12 developed economies by 2035, according to Accenture.
Already, AI is beginning to find uses in almost every industry, from chat bots that offer round-the-clock financial advice to helping doctors diagnose cancer. The technology is being used to power driverless cars, eradicate certain diseases, and better predict the weather and effects of climate change.
Insurance is an industry particularly suited to AI, as it involves lots of data and repetitive processes. In fact, the insurance industry has been an early adopter of machine learning, partnering with technology firms and investing in start-ups on a range of applications. Tata Consultancy Services' Global Trend Study found that of 13 industries surveyed, insurers invested $124m in AI, almost double the $70m cross-sector average.
Transforming the corporate insurance sector
According to Michael Bruch, Head of Emerging Trends at AGCS, AI is likely to transform the insurance industry, to the benefit of customers and insurers alike.
There is lots of potential for AI to improve the insurance value chain by making it more effective in addressing customer needs, delivering value on time and at lower cost.
AI is likely to make a big difference in three key areas. Initially it will help automate insurance processes, such as claims and underwriting. But over time it will give insurers and their clients a better understanding of risks, and it will change the way insurers interact with customers.
To date, insurers have mainly focused on developing AI applications for personal lines, but increasingly they are turning their attention to commercial insurance, including the large corporate market.
AI has the potential to bring about significant cost savings for commercial insurance, as well as speeding up the insurance transaction process and enhancing service in areas such as analysing submissions, checking or verifying policy documents, developing new insurance solutions and services or flagging up potentially fraudulent claims.
According to Michele Lagioia, an affiliate of the Italian Association for Artificial Intelligence, the claims process, in particular, would benefit further from increased automation.
“AI and automation would make for a much faster and more efficient settlement for lower value claims. Even with more complex commercial claims, AI could support claims decisions, speed up some processes and make for a more customized claims service,” says Lagioia.
Insurers in the life and health market are already using AI to review and analyze policy wordings and validate claims. Last year, start-up insurer Lemonade set a new world record when it used an AI claims system to accept and pay a claim for a lost item of clothing within three seconds, and without any paperwork. And many commercial lines of business, such as motor and workers’ compensation lend themselves to automation.
By automating repetitive tasks, people would be free to focus on value-added work, like client relationships, risk assessment or providing technical support.
The 24/7 insurer
AI also creates opportunities for commercial insurers to interact with their clients more frequently and more conveniently, as well as offer a more personalized or tailored offering. For example, in the personal lines arena, Allianz has developed Allie, an online assistant available 24/7 to answer customers’ questions.
“AI could be used to increase the points of contact with customers, ensuring an insurer is available at any time, and able to reach out and offer tailored products and services,” says Lagioia.
“For large commercial and corporate clients, insurance needs to be bespoke,” says Bruch. And while machine learning will support automated services it will also change the way insurers interact with clients and will allow for a platform approach to service. AI can help create an environment for insurers and third parties, offering a more targeted spread of risk management and insurance services.”
In addition to improving the insurance process and service delivery, AI could also boost data and analytics, the backbone of the insurance proposition, Bruch predicts.
“AI will be the key to unlocking data, especially as more data is made available by the Internet of Things (IoT). It could enable insurers to better understand customers’ risks and help clients reduce their risk, as well as find potential solutions for risks that may not be currently insurable.”
Allianz is already using machine learning to carry out risk assessments and to support automated underwriting in the small- to medium-enterprise (SME) space, a trend that will extend to the large commercial market. AGCS has developed a tool that uses machine learning to identify accumulations of business interruption risk in supply chains. The tool analyzes big data to identify and map networks of critical suppliers across industries.
“AI will support underwriters with the analysis of data and assessment of risk, helping to identify accumulations and price risk more reliably, while the insights gained should enable insurers to enter a more meaningful risk dialogue with clients,” says Bruch.
In cyber, for example, AI could play an important role in risk mitigation and risk assessment, benefiting both insured and insurer.
AI-powered analytics could help companies better understand their cyber risks, improve cyber security and even defend against cyber-attacks. At the same time, the technology could assist insurers in assessing cyber risk and spotting accumulations of cyber exposure.
“AI can improve commercial clients’ risk visibility. There are many areas – like reputation, cyber, supply chain and economic and climate risk scenarios – where machine learning could help companies better understand their risks,” says Bruch.
As the technology becomes more sophisticated, AI applications for analyzing risk will evolve, predicts Bruch.
“AI could act as an ‘intelligent agent’ able to create different scenarios and outcomes, and potentially take decisions. The next generation of machine learning will move from increasing risk awareness to decision-making.”
AI will also work alongside other technologies, most notably the IoT and blockchain, to increase our understanding of risk and enable insurers to offer new, faster and more customized services. For example, sensors on shipping containers are already providing data on the location and condition of cargo, which, once analyzed, can trigger insurance cover or mitigation measures if the goods are damaged.
“Insights gained from data and AI-powered analytics could expand the boundaries of insurability, extending existing products, as well as giving rise to new risk transfer solutions in areas like non-damage business interruption and reputational damage,” says Bruch.
AI brings new risks as well as benefits
In addition to improving the products and services insurers offer clients, AI will also impact insurers in another way – by introducing new risks to society.
AI technology is in its early stages, but it is expected to find a multitude of applications over coming years. AI is already making driverless cars a reality, but it is also leading to improvements in speech and image recognition.
However, like any disruptive technology, AI comes with risks and will have far-reaching implications for the economy, politics, mobility, healthcare, security and the environment.
AI could, for example, disrupt the labor market, changing the nature of some roles while eliminating others. Technology advances are also likely to increase the frequency and relevance of regulatory updates as governments and society struggle to keep pace.
AI will also need to be adequately tested to ensure it is safe before being introduced to the real world. For example, an AI agent for portfolio management could work according to specifications when tested in a development environment but then behave unexpectedly in the real world, making illegal investments. While AI could boost cyber security, conversely, misuse may also increase the risk of cyber-attacks if malicious hackers can train AI to attack.
For business, AI will bring reputational risks. Unintentional errors or unexpected consequences of AI applications could negatively impact consumer trust and cause reputational harm.
Liability and regulation is an area of particular relevance to AI risk. While AI agents are taking over decisions from humans, they are not yet legally responsible for damages caused by their actions. Ethical concerns are also likely to arise as AI pervades society, as will concerns for the accountability of AI and its ability to make transparent and auditable decisions.
Appropriate risk management strategies are needed to prevent, mitigate and transfer these risks and maximize the net benefits of AI in society. By addressing areas such as accessibility, safety, accountability, liability and ethics, responsible development and introduction of AI becomes less hazardous and the role of insurers will be key in ensuring risks are properly mitigated. New types of insurance coverage will help transfer and manage emerging risks.
 Accenture, Why Artificial Intelligence is the Future of Growth, 2016
 Tata Consultancy Services, Global Trend Study, Getting Smarter by the Sector: How 13 Industries Use Artificial Intelligence