And, just to round out the list, Netflix looks like one that’s actually trustworthy. They don’t advertise, and the info they have on me is pretty limited, anyway.
Netflix - I am conflicted between the choice of giving up my privacy to provide better suggestions. How about I actively go research for a particular movie on my own? Why do I need AI telling me what to watch. Word of mouth is a fantastic way to get recommendations. On the other hand, I would have not found interesting music on Spotify if it wasn't for their recommendation system / radio.
> Netflix - I am conflicted between the choice of giving up my privacy to provide better suggestions.
I'm confused by this conflict.
Netflix uses your personal data to help you get more out of the service. An argument could be made that they do this because subscribers then tend to remain subscribers for longer; but the counterargument there is that it is because those subscribers are getting more out of their subscriptions. I struggle to see any scenario where recommendations make people who are getting nothing out of the service stick around.
Apple uses personal data to make recommendations as well, but these are advertisements, because they are recommending additional products for purchase (and typically the products of third parties for which Apple gets a cut).
Is the difference here that Apple asks before they use this information?
> giving up my privacy to provide better suggestions.
The act of watching something on Netflix (or even just surfing the app and not watching anything) constitutes giving up your privacy. Netflix now knows (even if it forces itself to forget) something about you. I don't think it is problematic for a company to use this information to improve its services. I don't think the cynical view–that Netflix does it because it keeps subscribers around for longer–really matters here. I can't imagine a scenario where recommendations convince people to stick around who would otherwise need, except in that they provide a better experience.
The owner of a cafe I frequent knows many things about me and uses them to provide a better service. She will sometimes have my drink waiting for me because she knows what I drink and when I am going to be there. She knows I am not going to skip out on a bill, so she will make my drink out of order and let me pay later when the line has died. Etc. etc. I don't see Netflix recommendations as a whole lot different than that; just less personal.
I think sharing personal information with businesses parallels sharing information with individuals. If I tell a friend something and he starts blabbing it all over time (for free or for pay) or tries to use it against me somehow, we're going to have a problem. The same with Netflix. If they try to use it to manipulate me or they start selling it to third parties: problem. I want to control what I share, with who.
You make this sound like a question of honesty vs pretense, but "what they can get you to binge" reveals a different set of motives. It's sometimes a question of self-control.
I want to be healthy. I also want to eat junk food. What do I "really" want?
Imagine an online grocery store which noticed they could profit more by hiding healthy options from me and selling me more junk.
Would their AI be acting according to my "real taste"? In a sense, yes. Would it be serving my best interests? No.
If Netflix can entice someone to binge watch TV all weekend, does that mean that's what they "really" wanted?
Every AI serves someone's interests. It's important to ask whose.
And if you want to make an AI that truly serves your users (rather than exploiting them), give them explicit control. Stop with this "we know you better than you know yourself" nonsense. Let them say things like, "yes, I watched that movie / bought that candy, but I wish I hadn't. Please don't use that for future recommendations."
What you actually watch.
This is why I hate YouTube recommendations. It feeds my monkey brain.
It's a lost cause fighting advertising.
I think that’s the next move as soon as the technology will be ready.
I guess it doesn't need to be real time, so no super graphic cards are needed... or can it be real time? If a website (e.g. a hotel booking site) uses an ad network with "remarketing" (for this example let's say Google), it can tell Google you just looked for a hotel in Venice, and the next site you visit that has Google's banner ads will advertise "Hotel rooms in Venice!". Imagine looking up sweaters on a clothing site, flipping to Netflix, and an extra walks by that looks like you wearing a sweater that you looked at a minute ago...
https://ew.com/article/2011/07/07/how-i-met-your-mother-reru...
General advertising includes the sign on the door or exterior of the brick and mirter storefront, as well as billboards, magazine/newspaper ads, and product placement in media. It ranges from almost entirely informative (on the storefront) to annoying and off-putting (some billboards, poorly done product placement). This advertising is passive.
Individually targeted ads use extra information about the individual to cater the ad, and thus have an incentive to collect this information. This advertising is active, in that it is bidirectional.
Passive advertising not only still provides some benefit for much less harm, it's also hard to distinguish from purely informational materials in some cases.
It's worth distinguishing these to prevent bike-shedding over details that can be safely ignored when we formulate our arguments accurately.
Edit: s/targeted/individually targeted/ to make it more specific and accurate. That is the theme of the comment, after all...
I'm not sure why you would say that. e.g. There's a old silo building overlooking a bay near where I live, opposite a park, which has a huge billboard ad atop it the length of the building, which dominates the skyline. Calling that passive seems odd. It's aggressive, incredibly tasteless, and a symptom of a huge failure of civilization.
edit: And well, this is Sydney. Last night there was a rally at the Sydney Opera House to protest recent events: The head of the Opera House had refused to allow horse-racing ads (!) to be projected onto the sails of the Opera House. A well-known right-wing radio shock-jock got involved, and abusively insisted the ads be shown. Then, incredibly, the Premier of NSW (i.e. the leader of the state) forced the Opera House to back down, and show the ads. There's been widespread incredulity and outrage here about that! So, yeah, it is a problem.
They are annoying (but that's not inherent, it's entirely possible for someone to buy a billbaord and put up a great work of art or something), but they aren't a problem in the same class as the individually targeted advertisements that incentivize massive data collection and correlation projects and privacy violations.
> Then, incredibly, the Premier of NSW (i.e. the leader of the state) forced the Opera House to back down, and show the ads. There's been widespread incredulity and outrage here about that! So, yeah, it is a problem.
I think that's (possibly) a problem, but in a way entirely divorced from the problem of advertising. It depends entirely on the reason. If it was forced because the Opera house if funded with government money, maybe that plays into whether they can/should reject certain ads. I imagine that would play into the discussion in the U.S. (where whether you accept government money affects whether you can legally exclude certain classes of people, to my knowledge).
Group targeting is less of a problem because then they are working on averages and predominant group characteristics, which does not mean any one individual does correspond to those.
For example, a billboard may target an area with a message mainly meant for LGBT people, or for Hispanic people, but living or frequenting that area foes not mean you are LGBT or Hispanic respectively. There is still a level of anonymity there that protects privacy.
I've updated my original comment to make this more clear, so thanks for calling out the ambiguity.
If an ad intends to confuse their audience to think they're purely informational material, they probably should not be shown to anyone. That's a completely different issue, and Facebook's ad censoring step is supposed to deal with that.
It also only works when the advertisers know information about you, and the more they know, the better it works. This presents a perverse incentive to get as much information as possible, even if it's not in the individual's best interest.
To lay it out clearly, it is in an individual's best interest to have relevant ads shown if any ads are going to be shown. It is not in an individual's best interest to have personal, private or even possibly harmful information warehoused at different advertising entities. The tension between these two things is what we're discussing, and I think the harm (or possible harm) from having that personal information aggregated far outweighs the benefits to the individual for targeted advertising. That targeted advertising it also helps the company is of little consequence in this specific equation (that is, should I as an individual want this, but of course matters quite a bit in explaining the status quo).
Largely depends on who you are, I think. I’m not opposed to sharing any of my information with other ad companies
What if it's a disease you just discovered you have, and you don't necessarily have all the information yet? Should you be targeted before you've had time to look into it, much less come to terms with it?
What if you have a child with health problems or are part of a pregnancy that has problems? Would you want companies that might take advantage of the situation to be able to target you?
What if you've just been turned down for a job? Should a for-profit college with poor qualities be able to target you right afterwards?
There are many times when we find ourselves in an altered state of mind (even if slightly) where we are more susceptible to our baser emotions. Do we really need to engineer a system so companies can better identify when we are in these states for their benefit and to our detriment?
It's fine if someone wants to knowingly trade their information away for some benefit, but they should be able to see what they are trading away and make an informed assessment. What we have now doesn't really allow for that, and once it's out there, it generally has little to no value later (once you're marked as an ex-con, good luck getting that our of everyone's data sets).
If the ad is interesting and relevant enough (which is very rare), the next action would be researching the product online, and very occasionally I'd be super happy to know about it. It sounds absurd if anybody would make important decisions like deciding what treatment to get for health problems based on what people say in ads without doing proper research and consulting doctors.
Of course now advertisers are looking to ultrasonic beacons to get around the lack of targeting and data collection, but for the most part product placements aren't the malicious and disruptive advertisements we've become victims of.
A major branch of Russia and China's attack is the ongoing negative PR campaign against tech companies like Facebook, Google, and Amazon. Can these companies improve? Sure. But their infractions are incredibly minor compared to the dictatorships rising in Russia, China, Turkey, etc that are murdering foreign citizens as well as large numbers of their own. They are stealing billions of dollars from their own citizens and taking the world down a terrifying path.
Please reconsider the easy targets of tech companies and have the courage to face true evil in these dictatorships.