With every new day bringing new use cases, the risks and opportunities presented by generative AI (GenAI) are becoming clearer than ever.
At the crux of the balancing act of utilizing GenAI while mitigating the threats it can pose is the same factor behind its recent evolution – its accessibility. It is this which sets GenAI apart from other elements of artificial intelligence, according to Barbara Fernandez (pictured), head of Insur_space at MAPFRE.
The “conversational” nature of the technology is enabling users from any background to access super complex and sophisticated applications at their own pace and in their own natural language. The fact that you don’t need to be an expert to use the tool is a critical factor behind its rapid acceleration, she said, and this democratization is also why there have been so many surprising applications of GenAI and its capabilities to date.
Sharing some insights into MAPFRE’s recent report into how different scenarios can map out ‘the role of insurance in a society embracing GenAI’, she noted that it is not posing new threats but rather has the potential to escalate threats already present in other digital technologies. Take, for example, cybersecurity and fraud, she said - they’re not new challenges for the insurance industry but the democratizing aspect of GenAI is making fraud much easier to commit.
“We need to be prepared as companies but also as individuals,” she said. “For instance, today, if you’re sharing audio, you are providing data to synthesise your voice, and even 10 seconds of recording is enough… Companies in insurance also need to work on your protection but, at the end of the day, the first and last door [threats] enter through is the one you open. Everything is about awareness and having the right training, both as companies and individuals.”
While the cyber and fraud implications of GenAI are its main threats, Fernandez highlighted that other aspects of this tool are flying somewhat under the radar. One of these is related to mental health, she said, because it’s clear that there is a growing addiction to and reliance on digital technologies. Its accessibility means GenAI is making it easier for humans to interact with machines, which could advance this addictive behaviour.
“Again, GenAI is not generating this addiction but rather increasing it,” she said. “In a specific scenario, [a person] could create a relationship with a machine, if the processes are so sophisticated that you feel you’re communicating with another person… This is an extreme scenario which would have implications for a person’s mental health. But let’s not go to that extreme. Let’s look at where we currently are, which is that we are completely connected. And the easier your life is thanks to technology, the more you want to use it.
“This addiction is something that could generate some mental health issues. Other issues relate to critical thinking because when you’re asking for a summary or specific insights from information that you’ve shared with a machine, the issue is that human beings are lazy. We tend to think the things in front of us are correct, which is very high risk.”
The power of critical thinking needs to be preserved and nurtured, she said, particularly in the context of such highly accessible digital tools. Another potential risk for companies to bear in mind is how GenAI will impact how people interact with each other and companies, which could lead to the role of organisations shifting. In the future, the number of interactions with companies may be significantly reduced, which presents a risk to those companies which currently deal directly with consumers. They need to question what their role will be in a society that embraces GenAI – and they need to start questioning that now.
As to how insurance companies can walk the tightrope of opportunity and risk presented by GenAI, Fernandez cautioned on the need to take a measured approach to the technology, and not get blown away by the hype surrounding this tool. When any technology appears, she said, companies need to deep-dive into its capabilities and test them out.
Companies also need to face that there’s a growing gap between the individual employees who are training themselves to utilise GenAI effectively, and company-wide policies around GenAI use. That gap is growing, she said, and companies need to face this if they’re going to bridge that gap because you can’t stop people using it on their personal devices – rather, the focus needs to be on training employees to understand the tool and use it properly.
“The first thing companies that want to make the most of GenAI need to face at a corporate level is the need to train employees – and that’s reskilling all employees, not just the ones deploying the algorithm or the technical team – because everybody needs to understand how this works,” she said. “The second key thing is to recognise that GenAI is useful for a lot of use cases but not every single thing.”
Fernandez’s message is clear - insurance businesses cannot afford to be shocked by the hype around GenAI but invest and test to ensure that it’s applied only where it makes sense.
