In a fresh press release from Harvard, their Wyss Institute has created a new, AI-designed Xenobot that is capable of reproducing itself. This is astounding. That is a requisite for life, and from the video footage of how these bots work, it is eerily biological what they they do. They are shaped somewhat like a Pac-Man, and the original xenobot collects up cells available in the test volume, and forms them into copies of itself which are then “born” to continue the process.
While I am undeniably impressed, I’m also concerned for a couple of reasons.
- The Gray Goo Scenario
- There is a concern in general that something like this can replicate itself uncontrollably, producing a synthetic biomass that does two things: eat raw material and produce copies. All in an endless do-loop. Of course, the researchers say these xenobots are easy to kill, but will such devices always remain so? Can they mutate in someway to take on a more durable life cycle? Worse, could a researcher with low scruples or a demented intent decide to make a weapon out of these things?
- AI-Development Self-Bootstrapping
- These Wyss xenobots were developed via AI directed by the researchers, with some controls in place, and that’s good. However, what if the technology is pushed a little too far, and an AI development cycle results in a new species of xenobot that acquires new functions, either through algorithmic chance or an AI that overtly inserts those functions as a means to a conscious end?
One of my more woo-ish thoughts is that the Terminator franchise was prescient in describing our future, and in particular I have thought that the Terminator: Sarah Connor Chronicles TV show was canceled (terminated) because it dealt with informing on small doses of what is being planned. In general, the confusions and psychological stresses of a machine indistinguishable from human, those were major plot points, and given the hour-at-a-time format, a lot of tutelage was mixed into the action.
Congruent with the OP, though, is the show’s concept of “Grays”. Not alien creatures, but Skynet’s human collaborators, tasked with making terminators seem increasingly human. In a way, the researchers at the Wyss Institute may be the first Grays in the real world. Maybe they don’t realize the danger, even as they offer lip-service that they do.