Google, Online Platforms Told by EU To Explain Search Rankings (bloomberg.com) 89
Internet firms such as Google, Amazon.com and travel websites should explain how they rank search results on their platforms, according to European Union guidelines published Monday that could help businesses to increase their online visibility. From a report: The guidelines "set the standard for algorithmic ranking transparency," Margrethe Vestager, the EU's digital chief, said in a statement on the European Commission website. Online platforms should identify what factors their algorithms use when they decide to prioritize some results and declare when a prominent listing is paid for, according to the guidelines.
Sure OK (Score:2)
Re: Sure OK (Score:3)
This is not the EU.
At this point, that'a not an unrealistic wish anymore.
After all, pay *should* be for actual work. Not for a token that resulted from that work, that can be infinitely replicated at no cost, with no abiity to prevent anyone from doing so.
That only works for products. Matter-energy. Not structure.
For structure, only a service business model works. Money for programming work. Ordered programming work too. With the resulting progra being as free as the house that builder built for you. He isn
CORRECTION: I meant: "... not the US!" (Score:2)
Damn... I gotta go to bed! :)
Re: (Score:2)
That would actually be a very awesome thing for the EU to do. And it would be far more useful than this...
Unfortunately, I doubt the corporations would ever allow that to happen. I can hardly begin to fathom how many billions worth of lobbying money would go into the opposing side of such a proposal
The EU trying to encourage malicious SEO? (Score:5, Insightful)
according to European Union guidelines published Monday that could help businesses to increase their online visibility. From a report ...
Well, that's insane. Its specifically Not desired to have sites learning all details of the algorithm in order to spam their sites up to the top by gaming the metric, instead of creating organic content which their metrics against the web are designed to evaluate – there's a necessary reason for this secrecy. What's next.. a "Spam filter transparency" regulation for e-mail providers? Perhaps an "antivirus detection algorithms and patterns list" transparency rule to help ensure software companies can avoid their programs being labelled as suspicious.
Google already explains how to be visible... publish a web site up to current web standards including HTTPS with fast-loading pages that contain original high-quality content people want to see with frequent additions to that content.
Re:The EU trying to encourage malicious SEO? (Score:5, Insightful)
The problem is that Google's explanations are exceedingly limited for reasons you outline. Which leads to natural accusations of favouritism. Which Google doesn't help when it obviously does favourite some sources over others as a matter of routine, often to the point of absurdity. Social Justice nonsense being pushed to the top over science that debunks it that is hidden being a good example here.
If they actually have to publish the standards, this does two things. One is the negative you outline above. Other is the positive that they actually have to stand by the standards they publish, rather than put the finger on the scales behind the scenes for whatever reason they deem reasonable.
Re: (Score:2, Insightful)
Re:The EU trying to encourage malicious SEO? (Score:5, Insightful)
Fun part. This:
>Or are you one of those that think because women are biologically different, they are inherently incapable of "men's" work?
Is a "right wing talking point" because it makes a claim about inherent differences between genders, rather than making a claim that they are interchangeable and fluid.
The entire "right wing talking point therefore wrong" is in fact one of the major problems with google weighting today. Some right wing talking points are scientifically correct. Some are not. So is science a right wing talking point?
Ask google on some of those topics, and find out that no, science is "settled" on many topics which are "left wing talking points". A literal impossibility for science, as science is always improving and never settled. But it's "not settled" for right wing talking points. To the point where you'll find literal scientists debunking left wing nonsense de-ranked and often downright removed from search results on google today, provided they fail to align with specific political views.
Re: (Score:1)
I hadn't considered Google that extreme but maybe it is. Would you give us a couple specific search examples so we can see what you consider so egregious? Thanks.
Re: (Score:3)
The best part about this is that I cannot, specifically because of how opaque google's search ranking is. On aggregate, this is observable. But any specific search ranking is weighted in a way we cannot evaluate, making "give me a specific search" argument impossible to make, because there are no "specific searches that generate specific results" on any complex topic on google.
Instead there are searches that are weighted in opaque ways, and that are additionally curated based on who, from where and at what
Re: (Score:3)
So in other words 100% speculation on your part, thanks.
Re: (Score:3)
Observe the anti-scientific far leftist mind at work. He sees this:
>On aggregate, this is observable. But any specific search ranking is weighted in a way we cannot evaluate, making "give me a specific search" argument impossible to make, because there are no "specific searches that generate specific results" on any complex topic on google.
As a basis for transparency, which would remove this problem.
And then proceeds to ignore the first sentence, pretending as hard as he can that only the second one exis
Re: (Score:2)
Nice personal attack there, cunt.
Re: (Score:2)
Thank you. I like getting under the skin of bigots. Got physical scars to prove I'm really good at it with far right types from my youth. Glad to see it also works with far left, and shows just how much similarities you two have.
Re: (Score:3)
What a bizarre rant. The OP is clearly referring to the fact that it is difficult to study the Google ranking algorithm for the reasons given, but somehow it gets twisted through the prism of your mind into this strange, unhinged pseudo-science theory about how "leftists" are incapable of rational though, and therefore intolerant.
Re: (Score:2)
The fact that you failed to comprehend this very simple point in the exact same way/direction that the person above did is excellent evidence of my point. Thank you.
Re: (Score:2)
Projection is strong with this one.
Re: (Score:2)
The best part about this is that I cannot, specifically because of how opaque google's search ranking is. On aggregate, this is observable. But any specific search ranking is weighted in a way we cannot evaluate, making "give me a specific search" argument impossible to make, because there are no "specific searches that generate specific results" on any complex topic on google.
This is a cop out. I get what you're saying that different people get different results, which is a little bit true, but it's irrelevant. At the very least you should be able to give us searches that -- for you -- generate the sorts of skewed results you're talking about, and if we really get very different results you can provide screenshots. To make it a little better, do your searches in incognito mode, which increases the probability that we'll see the same results.
But I don't think you can find any e
Re: (Score:2)
>This is a cop out. I get what you're saying that different people get different results, which is a little bit true, but it's irrelevant. At the very least you should be able to give us searches that -- for you -- generate the sorts of skewed results you're talking about, and if we really get very different results you can provide screenshots. To make it a little better, do your searches in incognito mode, which increases the probability that we'll see the same results.
>But I don't think you can find
Re: (Score:3)
Well, if you are confusing biological sex with gender, then yes. Because science does not say genders, sex, and biology are always the same, nor does 100% of the population fit into nicely-defined categories. If a statement like that is being used to argue that gender separate from
Re: (Score:2, Insightful)
>Well, if you are confusing biological sex with gender, then yes.
Yes, that is the far left argument. It's the same argument as far left argument of "2+2=5". It's based on "if we redefine what two means, plus means, equals means, four means and five means, we come to a conclusion that everything is fluid and therefore two plus two equals five". It's a world play to deny science.
I'm not interested in far leftist redefinitions of words that enable this sort of subversion of science any more so than I'm inte
Re: (Score:2)
>Well, if you are confusing biological sex with gender, then yes.
Yes, that is the far left argument. It's the same argument as far left argument of "2+2=5".
No it's not - there are plenty of physical counterexamples in the world from your false notion that this is a political argument. For example, Trans women who were assigned male at birth may be biologically the same as many men in terms of genetics, and some anatomical characteristics at least until/unless they choose to undergo some surgery an
Re: (Score:2)
Another good example of just how incredibly damaging this attitude toward malleability of reality is:
>Trans women who were assigned male at birth may be biologically the same as many men in terms of genetics
It is false on the face of it to make such an incredibly stupid claim. Even identical twins are not biologically the same. And yet, as the demand for ignoring reality and science in the name of ideology requires the absolute certainty of these incredibly self evidently stupid assertions as if they're
Re: (Score:2)
It is false on the face of it to make such an incredibly stupid claim. Even identical twins are not biologically the same.
Well, now you are being deliberately obtuse and sputtering bad-faith arguments. Biologically the same does not mean identical. In this case we're referring to biological differences commonly associated with different sexes will not exist.
Re: (Score:2)
>Well, now you are being deliberately obtuse and sputtering bad-faith arguments.
You spent too much time looking in the mirror. You are now projecting your faults upon others.
>Biologically the same does not mean identical.
More redefinition of words to defend the indefensible.
>Again, your statements are extremely disingenuous here - there is no rejection of the core concept of science in these statements.
The pseudo-scientific field that is pushing those things, gender studies specifically rejects sci
Re: (Score:2)
Language evolves, deal with it. Gay used to mean happy, now it means homosexual. That's just how language works.
The first use of gender to refer to the social aspect was in 1955. Do you have an alternative word or phrase that you would propose to refer to this concept, or are you just trying to use Newspeak concepts to make it impossible to discuss it?
Re: (Score:2)
You mean "expression"? Yes, that word still exists. It's just that far left dogma has decided to pretend really hard that "gender" is the same thing as "gender expression" because it allows them to pretend that "gender" and therefore "sex" is fluid.
Exact same methodology as redefining racism as something that only some racists can do rather than a universal "discrimination based on race".
Re: (Score:2, Troll)
So your issue is that you want gender and sex to mean the same thing, and be immutable?
Well bad news, even biological sex isn't as binary or immutable as you like to think. The IOC really wishes it was, but after trying for a century they have been unable to come up with a clear definition that works for everyone.
Re: (Score:2)
>Well bad news, even biological sex isn't as binary or immutable as you like to think
And in other news, racism is privilege plus power, and white people are inhuman demons.
All of these views are direct product of the same language games played by far left bigots in their pursuit of power.
Re: (Score:2, Troll)
And in other news, racism is privilege plus power, and white people are inhuman demons.
This is why we can't have a reasonable discussion about this.
Re: (Score:2)
Obviously. When you dehumanize those you're talking to to that extent, discussion becomes impossible. And you as one of the most notable representatives of ideology that does this as a matter of routine mouthing this brings me hope.
But then, I suspect that you're not talking about the evil that is your ideology, but the fact that those that oppose it dare to speak against it.
Re: The EU trying to encourage malicious SEO? (Score:2)
women are biologically different
Excellent example. Posing that question to google yields the top hit as the Wikipedia article talking about X and Y chromosomes. Then there are several pages of results of articles mostly asserting that there are no differences between male and fmale brains. You don't start seeing anything talking about any other aspect of biology for about 6 pages.
Re: (Score:2)
I too would like an example of social justice results being favored. Perhaps the author would give an example?
Re: (Score:2)
Which Google doesn't help when it obviously does favourite some sources over others as a matter of routine, often to the point of absurdity. Social Justice nonsense being pushed to the top over science that debunks it that is hidden being a good example here.
Got any examples of this?
simpler (Score:2)
It's simpler than that.
The EU is just mandating that all of your algorithms are belong to us . . .
Re: (Score:2)
Except that of course, they lay no claim to ownership or copyright of relevant algorithms, so that statement is false on the face of it.
Re: The EU trying to encourage malicious SEO? (Score:2)
I don't think they are demanding the release of the recipe, only the ingredients to use a well known anology.
Re: The EU trying to encourage malicious SEO? (Score:1)
PROTIP: Site owners know and game the sysetm anyway, and have been for years.
Security through obscurity is a fallacy for a reason.
Good rules would not need obsucring anyway. Because the rules by definition would prioritize good sites.
If it needs obscuring, that means it's cheating, and using cheap tricks. Gameable tricks.
That means it's shit code and must be fixed.
Which is actually what this is all about, actually.
And I haven't even mentiones Google itself gaming the system! Which is also what this is all a
Re: The EU trying to encourage malicious SEO? (Score:2)
No, obscurity is definitely a legitimate means of security. Otherwise, surveillance would be pointless, we just jump right to the weaponization.
A search algorithm, no matter how advanced or unbiased, can be gamed. Remember that google originally did so well because they came up with the idea that more links to you means more popularity. The spammers just worked around that problem by link farming, so it's no longer useful. One metric google started using was measuring how many times people would click on a
Re: (Score:2)
Re: (Score:2)
Security through obscurity is a fallacy for a reason.
Its not a fallacy that obscurity facilitates security.
Sure. A flaw that is unknown to attackers might be eventually found out and exploited, so they should be mitigated.
However, a flaw that becomes known to attacks will always be exploited.
Furthermore, All systems have flaws - There cannot be a spam filter or ranking system on a search engine that is perfect and has zero weaknesses for spammers or adversaries to exploit.. It is an endless war betwee
Re: (Score:3)
The two alternatives are:
A) Companies have an even playing field (or as even as is practical to make it) and vie for attention.
Or
B) Google decides what matters and what doesn't, with no ability of anyone, especially you and I, to determine what criteria they use. And a history of abusing it for their own ends (with "Don't be evil" being the punch line of a sad joke), and a recent history of being increasingly political.
Re: (Score:2)
These are not good alternatives.
B) Google decides what matters and what doesn't, with no ability of anyone, especially you and I, to determine what criteria they use. And a history of abusing it for their own ends
Google deciding what matters and not has worked very well for the past 20 years, and its an Unfounded statement that Google "abuses it" - although they obviously tweak the rankings -- this is where it comes in handy to conduct your search using multiple competing search engines (Bing, etc)...
Re: (Score:2)
I *remember* when companies had an even playing field and could trivially manipulate (or hire someone to manipulate) the search rankings. It was terrible, and it was getting worse very rapidly -- the amount of time I had to spend hand-crafting search criteria was doubling every couple of months, there at the end.
And then Google came along, and they *fixed* it, by effectively eliminating SEO companies' ability to artificially manipulate the search results. And there was much rejoicing.
Re: (Score:2)
I *remember* when companies had an even playing field and could trivially manipulate (or hire someone to manipulate) the search rankings. It was terrible, and it was getting worse very rapidly -- the amount of time I had to spend hand-crafting search criteria was doubling every couple of months, there at the end.
And then Google came along, and they *fixed* it, by effectively eliminating SEO companies' ability to artificially manipulate the search results. And there was much rejoicing. And now, we live in a world where I can search using whatever criteria pop into my head, and the thing I'm looking for is *consistently* on the first page of results, and frequently "above the fold" (i.e., I don't have to even scroll down, to reach it). This is a *massive* improvement for the user.
If you don't remember the time before Google, you don't know the hell you've been spared.
That's not why I started using Google. I switched from AltaVista to Google because Google's results showed a snippet of text around your search terms. AltaVista just showed you the page title and first line or two of text. I found Google's version made it much easier for me to evaluate whether the page contained the information I was looking for or not. I did miss AltaVista's "near" search term operator :(
Re: (Score:2)
Google's results have been steadily getting closer to cat /dev/urandom. The majority of my searches return garbage link and text farms even with quotes wrapped around words. Sometimes those words are ignored even with quotes.
Re: (Score:2)
Re: (Score:1)
What you expose you can regulate, set methods and guidelines for search and search visibility, break those guidelines an pay penalties triple what you made, plus fines and repeat offenders custodial sentences. It is pretty easy to track, all of a sudden nothing companies reach higher up the search rankings. You do not need to audit all searches for validity just the top 10 sites of the top 10,000 searches, it would catch them out pretty quick. You can use computers to track even more the top 100 of the top
Re: (Score:2)
Google does actually give out hints about how to improve your search ranking. Don't cover up your content with overlays, don't have excessive ads, don't serve a different page to the search engine and the user, don't keyword spam, load fast, use HTTPS and network with other sites.
The EU has noticed that this advice isn't complete though. For example if you run a price comparison site you may find yourself struggling to compete with Google's own price comparison service, because for some reason it always get
What is the modern equivalent (Score:2)
of massive meta-data headers?
Re: What is the modern equivalent (Score:2)
Link farms?
Still working great, by the way. Even without the alogoritm being public.
Simple (Score:2)
Lossy node weight training involved? (Score:5, Interesting)
Algorithms that depend on "trained" weighted nodes or sequences (GA's) can be difficult to reverse engineer if the training set is lost or unknown. Humans are repeatedly asked to explain their decision process when results are contested or controversial, and many AI-based systems will probably want a similar functionality, as infrastructure and political info are increasingly impacted by AI (or at least statistics-based "training".)
Factor tables (see my sig below) are one approach that could help add regimentation to AI engines & teams, making AI more comparable to accounting rather than Doc Brown's ad-hoc lab.
Re: (Score:2)
Re: (Score:1)
Re: Lossy node weight training involved? (Score:2)
What else do you expect it to be?
Neutral?
There is only one single neutral function. And it is the identity function. And everything equivalent to it.
Ditto for the brain, actually. You cannot stay neutral and process some information at the same time. It is a literal neurological imposibility. Want it neutral? Remove the human. (And any media.)
And then your own senses and mind will do it anyway.
Re: Lossy node weight training involved? (Score:2)
Hint: Humans normally don't know why they made decision, and will make up a reason on the spot.
Or somethine worse, if they don't manag to rationalize it.
This is why asking humans about a crime you investigate, is bad criminology. Even if often common practice.
Re: (Score:1)
I will agree that's often the case. But at least one can blame a specific manager or professional*. Our legal and business habits know how to deal with that. A manager or engineer telling a jury "the bot told me to do X" won't fly.
* There are exceptions. Corporate mayhem often takes place in meetings where there is intentionally no written record, and if something bad happens, they all blame each other, saying, "That's
Re: (Score:1)
Nowadays algorithms are designed to put politics first and foremost at Big Tech. AI algorithms and datasets that produce factually accurate and politically inconvenient results are no longer tolerated. Unfortunately science has taken a back seat to politics and it is no longer possible to get quality information from big tech. Until this problem is resolved your field will continue to devolve and lose the trust of the public. The question is if AI can be saved from censorship and woke politics. Unless it is
Re: (Score:3)
Example? There is sometimes "social logic" to that. For example, giving statistics that "Group X commit more crimes" to justify discrimination and hatred usually cause more heat than light. Disenfranchised groups have almost always committed more crimes, because they feel they have less to lose and don't feel a sense of common community. That's human nature, not race. Without such contex
Re: (Score:2)
Re: (Score:1)
I'm opposed to any and all forms of identity politics, therefore I'm going to call out your presumption of racism, sexism or whateverism. I'm not trying to give you a hard time, it's just that you have illustrated the very issue that I was describing.
The presumption itself shows a bias and that bias is exactly the kind of issue that infects modern AI and creates the very issues of trust that I was referring to. These issues are why reporters and the media have almost the lowest level of trust in the US and
Re: (Score:1)
It was caused by a collection of things ginning up a storm. Us-vs-them racism has proven one of the most powerful motivators to get a population to fight and die for more territory and power. If other factors line up properly, then bigotry can light a big fire. Loss of factories in the US was one of the reasons T's tribalism got him in power. (The reason for factory losses is another discussion.)
Re: (Score:3)
This is part of what the EU is seeking to address. We don't want our lives to be ruled by algorithmic black boxes.
If companies want to use algorithms they need to explain how they work and show that they are now biased or unfair, and allow for human review on request.
Google Search Console (Score:2)
Google Search Console gives you quite a bit of advice on how to improve your search ranking and the factors involved.
Not only would reveling all the details be giving away proprietary information it would give people who try and manipulate search results even more power.
Breathtaking... (Score:2)
...that they even try.
Re: Breathtaking... (Score:2)
We're nit the US government, buddy. We don't suck coprorat e dick. Well, at least not non-EU corporate dick. ;)
Re:Nunya Damn Bidness (Score:5, Insightful)
It's none of the EU's damn business how Google ranks things in its search engine.
Bullshit, Google has a 90% search market share and if they are abusing that monopoly to play favourites and drive competitors into bankruptcy then that is sure as hell is the EU's business and the higher the fine Google gets handed the better. The only way these corporate assholes seem to learn is when they are handed multiple billion dollar fines.
Re: Nunya Damn Bidness (Score:3)
Right. Fuck off with your product sales and IP adresses and domains then.
Signed, the citizens of the EU.
Re: (Score:1)
They will never give that information, the search algorithm is Google's core business, it's cash cow. It is tuned to maximize revenue for Google. My guess is it is a ML backed bandit algorithm that learns which search results lead to profitable clickthroughs.
Re: And the result will be ... (Score:2)
Oh trust me: $500 million in fine plus $1-10 million for every day late in conforming to the rules will get them there. The EU did that before.
Youtube Buried My Channel (Score:2)
What I would like to see is ... (Score:3)
Explain the algorithm to the EU? (Score:1)
Wow, how dummied down will you need to get for a bunch of politicians to understand the whats and whys of Googles search. And when they get the 50,000ft view of it, they will make decisions on they're knowledge of technology? I think we've all been in those meetings. Knowing it's a waste of time. Knowing they will not understand. Knowing they already have their mind set to how they want it to work. But not knowing anything about how technology works.
Sounds like a winner to me!!
Swell. (Score:1)
It's all about fair play (Score:2)
The EU's GPD is about $18 trillion, while Alphabet's market value is ~$1 trillion. Therefore EU has to treat Google as a foreign country and have clear evidence that it does not actively engage in hostile financial activites by disfavouring EU companies in its algorithms.
Simple (Score:2)
1 ) is there a Wikipedia page for it? If so, put that 1st.
2 ) Otherwise, quickly have Alpha write a WP page, add that and then put that 1st.
3 ) Everything after the 1st result is based on what other people have looked at in the last 5 minutes, regardless of the search terms put in by this user.
Ta-da!
Seriously, Google's search is complete shit and DDG is following it over the "just serve what everyone else is looking at" cliff. I'm tired of being served WP pages first when there is an actual site devoted to