. . . or will it be fairly easy to build waterborne chaff to confuse the detectors?

Spam the shark, as it were.

If the effectiveness of the electric cables run in front of some Australian beaches is anything to judge by, this tech is pretty useless. Too trivial to jam.

You could sell it dodgy fin-ancial schemes.

Sharks stay away from the electric lines because they indicate danger/headaches to them, which are neither here nor there for the shark-in-a-box --- it is controlled by humans or programmed to ignore such strong and recognizable signals.

The presence of chaff-like stuff would indicate something's up, or indeed make the sharks go on a wild-goose chase and waste time. Here again, roboticists can develop chaff-recognition tech or soft where sharks are just stumped.

I'd be pretty optimistic something useful comes out of it, if only a sole-vacuum to precision-fish large but not pregnant flat fish. I'm in favour of research with edible by-products; memory research should be done preferentially on lobsters etc (given the continual need for `naive', unused, individuals).