I've been writing about this recently, the conclusion I'm coming to to is artificial intelligence is only as dangerous as the programming and / or programmers behind it, because at its very best it is automated decision making, if the programming is crap or biased and it often is then the outcome is crap or biased.
What I find deeply interesting and possibly a bit more disturbing, is artificial or non-human higher sentience (ie an ability to comprehend ones own demise) partly because it gets us possibly to questions of our own consciousness and sentience, and secondly the systems of restraint in something like that would be entirely outside of human control in the same way that one human is entirely outside of the control of any other human, bar evolved sociability inducing empathy and co-operation or through coercive means. Got onto this subject a while ago after reading that researchers at MIT and Cambridge concluded that the internet (or parts of it) might be already partly conscious.
I really dont see what issue there would be with environmentalists using AI any more than the AI used in twatter, or in Musk's crappy and rather stupid electric cars.