Artificial Intelligence Brings Risk Management Opportunities, Risks

Artificial intelligence (AI) can provide opportunities for organizations to improve their risk management but also may create some new risks those organizations must address.

Speaking as part of a recent webinar presented by the Federation of European Risk Management Associations (FERMA), Philippe Cotelle, head of insurance and risk management at Airbus Defense and Space, said that with the rise of AI and related applications, FERMA realized risk managers must familiarize themselves with AI and its potential impacts.

"It was pretty clear that AI was taking on more and more importance, and as risk managers we needed more awareness of the developments and to apply that awareness to the risk management process," said Mr. Cotelle during the webinar titled "Why Risk Managers Should Look at Artificial Intelligence Now."

Risk managers need to consider how to integrate risks generated by AI into the enterprise risk management (ERM) process and how to understand the scope of AI risks, he said.

To that end, in April FERMA released a white paper titled Artificial Intelligence Applied to Risk Management. The paper aimed to assess AI's value to improving ERM and to set out risk managers' role in helping their organizations' leadership understand the opportunities and challenges posed by AI.

The white paper notes that AI's ability to process large amounts of data and automate tasks could allow risk managers to respond faster to new and emerging exposures. And, by acting in real time with predictive capabilities, risk management could better support senior management decision making.

Mr. Cotelle noted that data is a critical element in any AI project. "There is no AI project without data, without quantity of data but also quality of data," he said. Risk managers need to identify areas within their organizations where they generate large quantities of high-quality data, he said.

Once you have the data, however, you have to be careful about how you use it, Mr. Cotelle said. "It all depends on the target that you have in deploying a new AI project," he said. Those projects might range from descriptive analytics that help determine why something happened to prescriptive analytics that can use data to guide future actions.

Integrating AI into the ERM framework involves five dimensions, Mr. Cotelle said: governance and culture; strategy and objective setting; performance; review and revision; and information, communication, and reporting. "All that put together will lead the way the AI framework is implemented," he said.

Mr. Cotelle outlined several potential benefits of introducing AI to risk management. Employing AI in data processing could increase visibility into risks, he said, while AI could also improve efficiency, reducing costs by automating day-to-day activities and improving understanding into correlations between risks.

AI could also help risk managers better manage complexity. It can provide increased awareness of new exposures, allow risk managers to provide better preventive risk advice, and allow near real-time responses in critical situations.

Artificial intelligence could also facilitate better business decisions, Mr. Cotelle said. "It will have a huge impact in terms of business decisions," he said. "This is the area that is perhaps most promising in terms of AI." Increased visibility and improved quantitative and predictive insights enabled by AI can help improve top management's decision making, Mr. Cotelle said.

As risk managers move toward employing AI, Mr. Cotelle encouraged them to "think big and start small.

"We need to encourage risk managers to develop a true AI road map," he said. Organizations need to develop plans for the infrastructure that will be necessary, addressing questions around build or buy, storage requirements, and available resources.

There are risks associated with AI, noted another panelist, Irina Orssich, team leader of Artificial Intelligence, Technologies, and Systems for Digitizing Industry in the Directorate General for Communications Networks, Content, and Technology (DG Connect) at the European Commission (EC).

Among them is the risk of organizations missing out on AI opportunities, Ms. Orssich said. She noted that AI's contribution to the global economy is estimated to reach $15.7 trillion by 2030. "There is a risk that we will be losing out ... in terms of being part of that development," she said, noting that only 12.3 percent of European Union companies used Big Data in 2019. "In order to tap into AI, we need to create the right framework," she said.

The EC, like other countries and international organizations including the Organization for Economic Co-operation and Development, the G7, the G20, and others are developing AI strategies, Ms. Orssich said.

Ms. Orssich said the EC is trying to develop a framework for identifying risks that AI might pose to individuals' fundamental rights and a risk-based approach to addressing them. Such risks might involve bias in the data sets used to train AI systems or flawed system algorithms, she said.

"When we have identified these risks, we would like to focus on high-risk applications," Ms. Orssich said. "We want to ensure a proportionate regulatory intervention."

The EC will seek to identify sectors where AI risks might be most prominent, such as automated transport, health applications, and financing, Ms. Orssich said, then look into the individual applications within those sectors. "Not everything in a given sector might be dangerous," she said, but there might be individual dangerous applications requiring regulatory attention.