I absolutely hate "smart" TVs! You can't even buy a quality "dumb" panel anymore. I can't convince the rest of my family and friends that the only things those smarts bring are built-in obsolescence, ads, and privacy issues.
I make it a point to NEVER connect my new 2022 LG C2 to the Internet, as any possible improvements from firmware updates will be overshadowed by garbage like ads in the UI, removal of existing features (warning: reddit link), privacy violations, possible attack vectors, non-existent security, and constant data breaches of the manufacturers that threaten to expose every bit of personal data that they suck up. Not to mention increased sluggishness after tons of unwanted "improvements" are stuffed into it over the years, as the chipset ages and can no longer cope.
I'd much rather spend a tenth of the price of my TV on a streaming box (Roku, Shield TV, etc.) and replace those after similar things happen to them in a few years. For example, the display of my OG 32-inch Sony Google TV from 2010 ($500) still works fine, but the OS has long been abandoned by both Sony and Google, and since 2015-16 even the basic things like YouTube and Chrome apps don't work anymore. Thank goodness I can set the HDMI port as default start-up, so I don't ever need to see the TV's native UI, and a new Roku Streaming Stick ($45) does just fine on this 720p panel. Plus, I'm not locked into the Roku ecosystem. If they begin (continue?) enshitifying their products, there are tons of other options available at similar price.
Most people don't replace their TVs every couple of years. Hell, my decade old 60-inch Sharp Aquos 1080p LCD TV that I bought for $2200 back in 2011 still works fine, and I only had to replace the streamer that's been driving it twice during all this time. Sony Google TV Box -> Nvidia Shield TV 2015 -> Nvidia Shield TV 2019. I plan to keep it in my basement until it dies completely before replacing it. The Shield TV goes to the LG C2 so that I never have to see LG's craptastic UI.
Sorry, just felt the need to vent. Would be very interested in reading community's opinions on this topic.
I think if enough people never gave them Internet access, the manufacturers would start adding in cellular modems to ensure they get the data flowing (that is, data on your viewing habits and sending you ads).
Having worked in this field, I can tell you how it usually operates: You want the most data for the least amount of investment. As soon as your operational costs start to eat into your already thin margins, the equation falls apart.
Complex solutions designed to capture data from that 1-3% of users who actively avoid it end up costing a lot more money than their data is actually worth. In order to make this particular solution work, you need to make enough money selling whatever tiny amount of data you get from those 1-3% of users to cover the cost of putting a cellular modem in all of your TVs plus the ongoing cost of paying various regional cellular networks to deliver that data to you. You are likely tripling or quadrupling the total cost of your data collection operation and all you have to show for it is a rounding error. And that is before we factor in the fact that these users likely aren't using the built in streaming apps, so the quality of the data you get from them is below average.
The cheaper option would be to set up an ad-hoc tv-to-tv network. You might not let your TV talk to the internet, but I bet your neighbour does, or if not, then their neighbour will.
The "Anti-Fraud Community Group" already thought of that:
https://github.com/antifraudcg/proposals/issues/17
...just in case you wanted to install ~~an ad blocker~~ malicious software, or something.
1-3% of users might not be enough people, but what is the break-even % of people to justify adding a cheap cellular modem? 5%? 10%?
You are likely not even doubling the cost of the data collection operation. We're talking under $0.50 in additional hardware per unit, with a relatively low data usage requirements. The servers to collect that data are likely already more expensive, and you can easily sell user viewing habits for way more than $1/month/user. You can use a prepaid low usage data-only eSIM with global roaming for less than $5/year, only renew it for the devices that don't get hooked up to a user's WiFi. If it was only needed for 5% of the users, or 1 in 20, you could still get a ROI of under a year.
With a device life of 5+ years, it's definitely much more than a rounding error. Keep in mind the profits go directly to the manufacturer, so it's a % of product cost in origin, not of MSRP... which is pretty much the reason why all manufacturers have jumped onto the data collection bandwagon in the first place.
That's what they do with CPAP machines.