Opinion | Man or Machine? Let’s Have a Chat
Recent praise for ChatGPT and generative AI has me scratching my head.
There are a number of amazing opportunities available through this tech— particularly for risk management and insurance. From processing millions of data points in a second to providing more accurate risk assessments from that data, there’s merit in turning to AI.
I’ve read the many industry pros in favor of it. But this is risk management and insurance. Why haven’t I heard anyone mention the dangers?
I’m not talking about I, Robot or Ex Machina-level dangers (though I won’t rule them out), but I am talking about data leaks, misinformation and other much more tangible threats that ChatGPT and generative AI could pose.
Did you hear about Samsung? Not once but twice did employees at the tech giant leak trade secrets by using ChatGPT. As one reporter from Mashable wrote, “ChatGPT doesn’t keep secrets.”
The Godfather of AI himself, Dr. Geoffrey Hinton — who was hired by Google in 2012 for the sole purpose of boosting its AI applications — quit his job in order to speak freely about the dangers ChatGPT brings.
His fear is that this tech has the potential to become more intelligent than humans and can pose the “existential threat” of people using ChatGPT and generative AI to create false information, deepfakes (like the spoofed image of Pope Francis in a puffer jacket) and more.
I’ll leave you with one final thought: Yes, artificial intelligence has its place in making society better. And yes, this industry should be looking at its capabilities to do just that. But as risk managers and insurance professionals, you have to take pause and run a risk assessment before trusting everything ChatGPT promises to be.
After all, do you trust that I wrote this? Or did I edit something ChatGPT threw together? &