Overall costs is the important aspect. I think very much for the time being going forward that the plan is many hundred millions use A.I. for entertainment or similar, enabling it to be affordable on big projects. I expect this given how A.I. now being asked to do services that were performed just as well before A.I. - primarily as a search engines. Frankly as someone as an extremely heavy search engine user up until 2015 and tapering off with DDG till approx 2017, both Google and DuckDuckGo are ghosts of what they once were.

Yes the other issue many would have, even with smaller low compute is it continues the path of subscription based services. I hear for example citing its use as a word processor, correcting grammar but surely this is a function that a decent A.I. could compute the definitive buy once ultimate guide program that can work offline and with less issue of leaking creative thoughts and ideas if a person was worried about such things.

Most western countries are surveilled by the likes of agencies similar to NSA - it happens and shouldn't be of much concern to the average punter, but A.I. is mostly likely going to be used by local, state and federal authorities to constantly monitor and with little overview, perhaps pandering to unknown interests or those with an axe to grind who wish that relevant interventions occur with those opposed to their grand plans. By interventions, it could be dire but even as benign as simply burying a perceived thorn under a heap of paperwork / legal wrangling / on time sensitive matters- never got a reply, if registered mail - oops oh your reply was lost ... etc.

Surveillance is probably one of my biggest concerns with AI. The ability to monitor millions of people isn't something new, but AI makes analyzing all information drastically faster and more efficient.

And what about the BIAS we mentioned earlier in the thread? Wouldn't that influence the surveillance? isn't that a bad thing? there is just so many things to consider.

Very much there will be a bias depending on who's supplying the product ... eventually I think we'll see attempts to ban use of foreign A.I.s especially in regard to the average user. There is IMO a good potential western or US based will probably dictate guardrails/ morality / TOS much like when the US (a bit before but around 2008) effectively banned anonymous people giving blow hard idiots a hard time, the peculiar outcome of an inquiry due to or because Facebook had a piss poor moderation strategy to maximise profits - nearly all the forum software being US based had to change their TOS as did hosting sites. The US these days seems to bend to readily to those who have money, sure money has always been part of the problem but nothing as blunt as it is presently.

As for my own personal fears, is the possibility, however slight, popularising the adoption of a personal A.I. for all internet users by way of a scheme that it will be deemed essential, like a personal web guardian which acts as interactive security and key pass, no A.I. buddy, no or extremely limited net access - essentially a minder. If that doesn't transpire, I worry that in 20 years or so, there will be moves towards devices which are just enough to run a connection to their A.I. guardian which does everything that is deemed acceptable for the user / allowed to do.