Neither, but the Chinese less. I don't trust anyone. Hardly a month goes by without some information breach making the news, some of them in governments.
Too much information is being collected, much unnecessarily. (Partly this is a hangover from migrations to computer-based systems, many implemented during a long era in which privacy concerns were almost absent and the general approach of analysts was "hey, let's collect all this stuff".)
Hayek explained the knowledge problem confronting would-be technocrats (mainly, political progressives). A couple of hundred years of events have shown that the more complex society becomes, the more things need to be decentralized to function well (efficiently, productively). Some parties in government make government more complex to attempt to match societies grown more complex, but they are falling behind - the gap almost monotonically increases. Thus much of their complexity is needless; thus much of the information they collect is needless.
The prospect of AI is a life-preserver at which they're going to desperately snatch (if they haven't already), hoping it will close the gap. It might, a little, but it too will fall way short. (Among other things there will be too much garbage input, and too much output which results in extraordinary claims based on sparse inputs.) But the potential abuses are so vast that no liberty-minded person should be willing to make the trade-off. Some will imagine and describe some saintlier version of Mark Carney wielding the power ("think of all the good they can do!"); eventually the apparatus will fall into lesser hands (as so many people lament about the powers wielded by the Trump administration). Looking at the mediocrity of people in general, we're always more likely to get the latter than the former.