Google's Video Nasties
bloomberg.com
bloomberg.com
From what I understand of online targeted advertising is that it's a way for corporations to target the viewer, as opposed to the video they are watching. I have a diverse array of interests and watch a whole bunch of videos on YouTube on a variety of subjects. I am not defined by a single video or channel.
As per my understanding, I am the commodity. If I am a history buff and I want to watch an Adolf Hitler speech for academic reasons or just curiosity, by casting aspersions on the channel or the video, you are effectively casting aspersions on me! Is the value of my time tainted because of one video or channel?
In my experience, when I look at something on Amazon, I invariably see an ad for it later on on some other site. It's clear to me that this is a connection between Amazon and me via the ad network. It puzzles me as to why there seems to be this outdated mentality about how this works, which incorrectly associates the ad with the page instead of the user.
I accept that the solution is difficult, but Google designed and benefits enormously from this system, so it's reasonable to expect them to do something about it.
Second, this is a (needed) power play on the part of ad buyers. Google has growing control over huge swaths of online and mobile advertising. Ad buyers need to push back, questioning the value of this sort of advertising and demanding more control over where their ads are displayed. Why should they trust Google to do the job of deciding where their ads belong for them? Google has proven time and again that they trust their own AI and machine learning tools far more than they should. And the longer they wait, the worse it's going to get.
If advertisers have no control over which content their ads appear near then they can legitimately claim, when criticised, that they are unable to pick and choose.
By giving advertisers control, when criticised, the critics can use that same control in order to pressure advertisers to limit ads to inoffensive content only (which, frankly, is almost nothing anymore).
But with the direction "being offended" has taken (i.e. someone on twitter will take offense to anything, and they feel they have a legitimate right to never feel offended) it won't be long before most content is inaccessible.
> (i.e. someone on twitter will take offense to anything, and they feel they have a legitimate right to never feel offended)
Your post was stronger without this straw man. Do you think a real person would say that they have a legitimate right to never feel offended? I don't think so.
That part of it is easy to mitigate by the auction process. I.e., the mean efficacy of ad placement is reflected in the price of the winning bid due to the usual market forces.
The less quantifiable part of it, "my brand must never appear alongside X," cannot be dealt with in the auction because the recall has to be perfect and it's not feasible to guess how the advertiser would decide all the tricky cases.
While you might be able to automate the process somewhat with ads that aim for direct conversions, it's still extremely difficult to automatically measure the effectiveness of brand advertising.
Advertisers will respond to the masses, not the most outraged hecklers.
Twitter seems to be proving that wrong these days.
When the same social circles are both in PR/HR and the offended class they will consider the same things valid criticism and may go as far as flying in the face of 90% of the quiet consumers of the product's interests.
I dunno. I find Nazis, cat torture videos and pro-ISIS propaganda offensive. But you are right, clearly free speech must be protected at all costs and even made profitable, and clearly it being remove from YouTube and no longer subsidized by major brands advertising would be a huge set back for it.
Maybe people could go to NaziCatTorturers.com and pay to see it or something.
That seems forced. I understand it's a link to another article that they wanted to include somehow, but it's specifically not hubris if they have a valid reason, and as just stated it's because it reduces inventory and adds to their costs.
http://teneightymagazine.com/2017/03/19/youtubes-restricted-...
They've walked it back with "our algorithm went overboard," but I wonder when people are gonna stop buying this excuse. If the automated system you put in place serves extremist ads or deems a whole class of people as objectionable, that's on you.
The algorithm works in this instance. His speech is hate speech by most definitions of the word. But that's the main problem. Hate speech includes reasonable hate!
If I had it my way, reasonable discussion about a culture should never be blocked. Even if it's hateful. Hate should not be banned, otherwise people will vent their hate in worse ways.
Here's his channel if you're curious: https://www.youtube.com/user/sotomayortv2
Example video: https://www.youtube.com/watch?v=ycYDUCGi7no1
In practice, and everyone knows this, filters fucking suck and either have terrible false negatives and don't block awful content, or terrible false positives and block stuff that shouldn't be blocked, or both.
That's nothing to do with this, which is that advertisers don't want their ads next to people saying rape is ok.
I don't think those restrictions are applied automatically. I believe they are set when a video get lots of reports. Maybe it only takes a few reports to restrict the videos, or maybe a human has to evaluate them before flipping the switch. As much as filters suck (I agree with you), I think a few false positives are worth it if we don't show a kid a video of some idiot saying rape is ok.
But I don't use youtube much, so I might be wrong... that's why I ask. :)