I wonder why most of the comments take it very casually and say may be issue with digitizer/ghost touch. Had it been any other OEM, this would have been such a big issue with anecdotes of why people trust apple products.
If you were genuinely remotely hacking a smartwatch, you’d be executing background processes to exfiltrate data entirely invisible to the user, not doing some bizarre remote desktop thing randomly tapping around on apps. The claim doesn’t pass the sniff test irrespective of the manufacturer.
From experience, you really can't just take user reports at face value. There's almost always something there, but it may or may not be what the user thinks it is.
So it's a good idea to apply Occam's razor.
Digitizer/ghost touch is probably the simplest explanation.
The only thing the hacked/pwned idea has going for it is the "We are in control" message, which is still a bit marginal if the watch really was hacked. (None of the other posts mention this and why would a hacker type that message in? Could be because it's a practical joke or maybe part of a phishing attack, but those are tenuous and nothing else mentioned supports those.)
If you have access to control the touch interface or to type a message on the screen you already have full control to the device. Look at the video, the input is so random, it's a software bug.
iOS developers who have done work for watchOS know that, during development, you are barely able to connect an Apple Watch to Xcode. It is flaky as hell. So thinking that someone can remotely control an Apple Watch is a bit hard to believe.
Ghost touches and typing "we are in control" with predictive text, sounds more plausible but still raises an eyebrow.