According to IBM research, ChatGPT is already capable of writing convincing phishing emails, raising concerns about the potential for hackers to exploit AI tools for cyberattacks. In an A/B testing experiment, 11% of employees fell for a ChatGPT-written email, while 14% fell for a human-written phishing email. Although ChatGPT lacks emotional intelligence, its ability to generate phishing emails quickly is worrisome for cybersecurity. While safeguards are in place, social engineers have found ways to work around them.
Insurance companies are increasingly using AI to combat fraud, with nearly 60% already utilizing AI for this purpose. The rise of deepfakes and manipulated images has posed a new challenge, prompting the development of software tools to verify the authenticity of images. While current AI tech delivers fraud alerts, developers are also exploring generative AI systems to assist investigators. However, challenges remain in terms of data availability, regulation, and the need for transparency and explainability in AI systems.
Why do we care?
AI good and bad at the same time. Considering AI-written phishing emails are nearly as effective as human-written ones, MSPs need to ramp up education and detection measures. It’s not just about installing spam filters; it’s about continuous education and advanced monitoring tools that can adapt as fast as the threats do.
The increased use of AI by insurance companies to detect fraud is a two-edged sword. On one hand, it offers an additional layer of scrutiny that can be beneficial in assessing risks and detecting fraudulent activities. On the other, the challenges of data availability, regulation, and transparency still need to be addressed. It’s why offering not just technological but also regulatory expertise is the critical insight for value.