I’m speaking of online data harvested through apps, websites, hardware (such as phones/streaming devices).
I mean if multiple versions of the same harvested data are being sold, wouldn’t the value decrease because of the competition? When it comes to aggregate data, how much financial value can there really be in knowing that a million office workers just clicked on the same cat meme?
How does the quantity of time and expense toward “personalization” not simply overshadow the return, given that no one can click on even a small percentage of those numerous ads, let alone buy the shit being advertised?
It just seems like there would come a time when the value of user data is sucked dry, or at least significantly decreased.
Seeing how 90% of the ads I get are reeeeeally missing the boat, they might already be over- harvesting my data?
Exactly.
Alternately, perhaps the algorithm knows me better than I know myself. Perhaps the humidifier I just bought is about to kick off a new hobby of collecting humidifiers. /s
Yep, just like a couple years back when I bought a part for my Toyota. Maybe I was about to collect the same party for a Mazda and a GMC
I almost spit out my tea, thanks!
It’s getting weird. I’m suddenly seeing a lot of things pretending to not be ads but specifically referencing my age or things generally blanketed to my age group. But weird things like life insurance and dick pills. Stuff I’ve never searched before because I already have employer provided insurance and my shmegan works as intended.
I think we could run in the same effect as energy have, when a new source of energy is found, it’s not reduce any energy use, it actually increases the use of it because is now cheaper.
With more data, more companies that use statistical analysis to processes it are going to pop up, new uses for that data is found and more data is going to be needed.
I work in insurance, with insurance bonds, nowadays the underwriters check not only the financial information of the company buying the policy, but also the personal and financial information of the executives, directors and their families. We didn’t went so far 10 years ago because it was expensive, in the future, maybe we even add employees and their families on the underwriting assessment too.
Yeah, I could see the financial value dropping, with businesses less willing to pay as much for harvested data, but I don’t see a point in time where they don’t attempt harvest every last piece of data on the off chance somebody wants it though. Advertisers paid insane amounts of money for targeted information, but even Google’s seen a huge contraction in their advertising revenue.
Doesn’t mean they aren’t frantically trying to harvest data more aggressively (just recently tried to bake it into the internet itself), just that our data is getting cheaper.
As AI generated word salads start polluting the Internet more and more, It will become kind of useless for training future AI models and advertising. It’s logical to think that confirmed human generated content would become more valuable perhaps… Check out the dead Internet theory
The cool part is that we’re definitely going to see people paying top dollar for certified human generated data, that happens to be generated by an AI that can beat Captcha better than humans can.
It’s a tiny bright spot in this mess, but I’m really looking forward to laughing at someone for it.
Hm I wonder if generative AI could somehow be used to create fake user behavior to obscure my real behavior.
Slight spoiler for Neal Stephenson's book "Fall (or: Dodge in Hell)".
That idea is used early-ish in the book to muddy the waters and regain some privacy for a significant character. There’s a subplot involving a conspiracy theory being intentionally spread through online message boards and this character becomes a target of harassment. By flooding the sites with a full spectrum of scandals and red herrings, they’re able to drown out the pieces of legitimate personal info and confuse people enough that they move on to new topics.
There already are tools out there that obfuscate your data by making it purposefully wrong, e.g. https://adnauseam.io/ is an ad blocker that also clicks every single ad sent your way.
Also benefiting the ad companies who charge per click, while not ever leading to a sale.
Data analysis, big data, etc. Already hit a point of data pollution concern. There are more datasets available than there are possible scientific and financial uses in a given timeframe. Basically, there’s people whose job is to sit around all day and try to find uses with any ROI for data that companies already paid for or already collected without any particular goal in mind. Because if they don’t, then those expenses will turn into losses.
Think of it this way, while companies would pay dearly for datasets for advertisement, for example, these include the data of several million people. While the individual data point you personally generated would amount to fractions of a cent in value. So, our data is already heavily devalued.
At the same time, the highest cost is in scrubbing, cleaning, tailoring and analyzing data. Turning data into actionable information. Collecting the data itself is only a small portion of costs. While the vast majority of data collected is low quality or straight up useless. To mitigate data pollution, there’s a whole field of data science whose sole job is to decide what needs to be deleted and thrown away, for it harms the usability of the rest of the dataset.
Equally, companies pay millions in online advertisements, but the individual impression you see on a webpage or search results cost a few cents at most.
I think the sources of data will change. Things like browsing habits will become useless, but other tools you use will harvest very specific data.
We’re already there. There’s a reason companies keep pushing this idea that every single goddamn thing in your household needs to be a “smart” device that’s connected to an account.
Well we’ve been on this trajectory for a long time but we’re nowhere near “there”.
Imagine a personal AI assistant. Not the bullshit voice assistants we have now, but an AI that is trained to assist me personally and predict my needs. Not only would it contain a wealth of information about me, but it could feed me promoted selections as well.
Nice idea, as the data becomes worth less, there will be more incentive to collect more data and more invasive data.
Vague hopes that it would drive advancements in compression methods and storage costs.
