It’s interesting to see the recent concerns around ChatGPT and AI as they are very valid, and I fear that too little is happening too slowly from a parameter standpoint. Fast forward to today and look at the neck-breaking speed of evolution of these wonderful and quite dangerous tools since I began ringing the alarm on the dangers of deep fakes back in 2019.
The tools are very entertaining and capable of producing believable images, videos, and audio of anyone without the need of any technical ability. Recently, a deep fake track of #Drake was posted online that was believed to be a new hit by many of his fans! A well-known vlogger in NYC by the name of Casey Neistat even used ChatGPT to generate a script for a vlog that sounded like something he created. There are a good number of free/low-cost tools that allow anyone to generate them. There are even tools to generate deep fakes of nudes and pornographic material. This is extremely disturbing!
My concerns go beyond the ease of use and accessibility of these tools and the destructive consequences on industries, society, and individuals by the tools themselves. We also need to concern ourselves with what data/images are being uploaded to these tools, who owns the tools, what access and retention controls are on them, etc. If a bad actor were to gain access to your voice print, likeness, and biometric data, what evil can they perpetrate? I argue that it’s almost limitless!
As far as regulation and governance, that ship has sailed over the horizon! The next enabler that will hit us is quantum computing. As quantum computing becomes democratized and accessible by the masses, how much worse will these threats and the damages coming from them become? Regulation and governance without the ability to detect and enforce rules are futile. We need to focus on these threats and ensure that we build awareness and tools to limit risk immediately, because it is here.
There are limitless benefits for society in these tools and at the moment, there are equally limitless risks. We need immediate recognition and awareness of the dark side, as well as a clear path to contain the beast while it’s in its adolescence.
My challenge to the good folks out there: we need to focus on detection tools and methods for cleansing the problem when it hits.
Originally published on LinkedIn, this post was written by Dr. Tomislav Mustac, Senior Director and Head of Systems, Cloud & Biomed Security, Mount Sinai Health System. The opinions expressed are his own and do not necessarily reflect those of his employer.