Fable 5 Shutdown
The US government has shut down Fable 5. Here’s my thoughts.
It’s probably not about the models. It’s probably about everything that the models can touch.
The real security vulnerability is a more-than-tacit realization that critical infrastructure throughout the United States is woefully unprepared for attack by malicious actors. Decades of software-supported infrastructure around power, water, transportation, etc. has been put into daily use. At the same time, that supporting software (and the networks that connect these systems) has not had anything near the level of security hardening required to affect the probing and infiltration efforts that AI-augmented attackers can now unleash. Thus, this is a concern affecting national security for the United States.
But, the argument goes, this is not a problem unique to Fable 5. Other models can theoretically be employed to undermine the US infrastructure. True, but Fable 5 is on the leading edge, and the US government has to pick someone as the first “poster-boy” to haul to the carpet. The other models have “left the barn” a long time ago. It’s not about “Why did you pull me over when all the other cars are speeding?” It’s about the fact that the authorities can’t pull everyone over, so they’re going to focus on the fastest (or, in this case, the most recent) on the freeway. Part of the hope is that the flashing lights will warn other motorists (Open AI, etc.) to SLOW DOWN.
There is no way the US government can simply be open and say, “We’re woefully unprepared.” They have to couch it in terms that portray the danger lays in the AI model itself rather than the rickety network of national necessities which allow us the illusion of prosperity, safety and security.
I’m only a few minutes in to listening to Nate Jone’s take on the issue (see The End of Unrestricted AI: Why Claude Fable 5 Was Just Forced Offline), but I had to stop and write down my own thoughts. And for those who are concerned that jobs are disappearing because of AI - just think of all the systems that need to be hardened and protected against AI-augmented attacks. There will be a lot of work to do in the coming years to shore up our defenses, and that will require a lot of skilled workers. We just don’t have enough Red and Blue Teams (let alone Purple Teams) to do the work.
Back to my video watching…