Google search ranking is editorial in nature and qualifies for First Amendment protection

A Google-commissioned legal paper on the constitutionality of regulating search results concludes that the such a regulation would violate the First Amendment. "First Amendment Protection for Search Engine Search Results" was written by eminent legal scholar Eugene Volokh and attorney Donald Falk, who argued that search-results are like the table of contents in a magazine, reflected protected expression in the form of editorial judgment.

In the case of a magazine, the articles are selected by a human editor. In the case of Google, the search results are selected by an algorithm, but the algorithm is created and managed by engineers who apply editorial judgment to the results. I absolutely agree with this conclusion.

However, I'm surprised to see Google in accord with me on this one. In all my discussions with googlers on this subject to date, I've always been told that search-results represent a kind of abstract "relevance," not anything as sticky and human as "judgment." It was as though Google's sorting algorithm provided a wormhole from the walls of Plato's cave straight into your browser.

Up until now, all the arguments against regulating search results I've heard have turned on this notion of search results being untouched by human hands. The reason that an unflattering "sucks" site appears at the top the search for a company's name is that the offending site is "relevant" according to some infallible mathematics of significance. To order Google to rearrange its search results is like ordering a parachute company to change the constant it uses in calculating gravity.

I've always hated this argument. Google regularly "tweaks" its ranking algorithm to provide "better, more relevant" results. These tweaks' success are measured by how "right" they appear, both to Google and to its users. They are, in other words, judgments.

I think that the editorial right to exercise judgment is much more widely understood than the sacred infallibility of robotic sorting. I certainly support it more. But I wonder if Google appreciates that it will now have to confront people who are angry about their search rankings by saying, "I'm sorry, we just don't like you very much" instead of "I'm sorry, our equations put you where you belong." And oy, the libel headaches they're going to face.

Here's Timothy B Lee reporting at Ars Technica:

The authors argue that this selection process is no different, constitutionally speaking, from a newspaper editor selecting wire stories to run, a guidebook deciding which attractions to feature, or a parade organizer choosing which floats to include. The courts have ruled that all of these editorial processes are fully protected by the First Amendment.

Moreover, the paper argues, the courts have held that First Amendment rights generally trump antitrust law—something of increasing concern to a dominant company like Google. "Antitrust law cannot be used to require a speaker to include certain material in its speech product," Volokh and Falk write. They point to a 1945 case in which the courts found the Associated Press had violated antitrust laws, but stressed that its ruling did not "compel AP or its members to permit publication of anything which their 'reason' tells them should not be published." Newspaper editors have the right to decide which stories should be included in their newspapers and which ones make the front page. This suggests that Google has similarly wide discretion to decide which links and other content will appear, and in which order, in response to any given search query.

Here's a quote from the paper itself:

In this respect, each search engine’s editorial judgment is much like many other familiar editorial judgments:

* newspapers’ daily judgments about which wire service stories to run, and whether they are to go “above the fold”;
* newspapers’ periodic judgments about which op-ed columnists, lifestyle columnists, business columnists, or consumer product columnists are worth carrying regularly, and where their columns are to be placed;
* guidebooks’ judgments about which local attractions, museums, stores, and restaurants to mention, and how prominently to mention them;
* the judgment of sites such as about which stories to link to, and in what order to list them.

All these speakers must decide: Out of the thousands of possible items that could be included, which to include, and how to arrange those that are included? Such editorial judgments may differ in certain ways: For example, a newspaper also includes the materials that its editors have selected and arranged, while the speech of or a search engine consists almost entirely of the selected and arranged links to others’ material. But the judgments are all, at their core, editorial judgments about what users are likely to find interesting and valuable. And all these exercises of editorial judgment are fully protected by the First Amendment.

That is so even when a newspaper simply makes the judgment to cover some particular subject matter: For instance, when many newspapers published TV listings, they were free to choose to do so without regard to whether this choice undermined the market for TV Guide. Likewise, search engines are free to include and highlight their own listings of (for example) local review pages even though Yelp might prefer that the search engines instead rank Yelp’s information higher. And this First Amendment protection is even more clearly present when a speaker, such as Google, makes not just the one include-or-not editorial judgment, but rather many judgments about how to design the algorithms that produce and rank search results that — in Google’s opinion — are likely to be most useful to users.

Scholar: regulating Google results would violate First Amendment


  1. So, if I create an algorithm that can create an near infinite amount of plots, I can claim copyright on those?

    And the true chess masters are the programmers of Deep Blue?

    1. You’d have to go to the trouble of printing or publishing them somewhere, and proving a level of originality.

      To your second point, I’d say yes.

      However, both of your points are utterly unrelated to the First Amendment.

      1. a) Why would I ned to publish or print them? Both a not necessary for copyright in most jurisdiction. Also, publishing could be in form an website anyway.

        b) Swell. I’ll get my exoskeleton  and register for the olympic games. 

        c) They are related to the argument that the First Amendment applies because it applies to persons who wrote the algorithms.  Even if said persons have no say to how the algorithms are being applied.  

        1. a) No printing nor publishing are a necessity for copyright but there has to be some kind of tangible form. You can’t have the idea in your head, nor as a possibility in the depth of a computer, ready to be written by your algorithm. So writing the algorithm itself wouldn’t suddenly give you copyright over “near infinite plots”. However if the plots were being spat out onto a website, you might have a case for the plots that have been published at least. However, you’d also have to show a level of originality, which I imagine would be the hard part. It’s one thing to program a computer to play a game of logic and skill. Its another one to program it to write creatively, originally and sensibly.

          Another way of thinking about it could be this: if you wrote an algorithm that generated and printed 3d sculptures or rendered amazing digital art would you own the copyright? Yes, of course you would. You’ve created everything that led to that outcome.

          b) Apples ain’t oranges. The Olympics have pretty strict guidelines about the fact you have to perform without physical or chemical enhancements. 

          Chess, well that’s a game of logic and skill. If logic and skill went into programming a computer that beats humans, I think that makes the programmers the masters. It doesn’t necessarily follow that those people should be allowed to enter their computer into every chess championship in the world.

          c) Firstly, neither Copyright, nor the question of who are the true chess masters has anything to do with the First Amendment. Copyright is a completely separate legal construct that must be interpreted on its own merits and the chess masters question, well I’m just not sure how you think that’s related.

          Secondly, the people that write Googles algorithms do have have a say in how they are being applied. They are writing them for the exact purpose of the algorithms being applied how they want them to be.

    2.  “And the true chess masters are the programmers of Deep Blue?”

      Of course not.  Deep Blue is programmed using many basic rules of chess, that all of the professionals know.  What makes a chess master is the ability look forward and choose the best path based on probably outcomes.  Writing a computer program to do this bears no similarity to being able to do it yourself.

  2. The programmers are Google are paid personnel.  Like Kirby back at Marvel.

    Also,  lets say I copy the  program  for 3-D sculptures  and run it myself, on my machine. Who’s the creator? The original programmer?  If yes, why isn’t it acceptable that Apple and others limit what people do with their algorithms? 

    1. Or the editorial staff at a newspaper, where the First Amendment applies routinely!

      Well in that case you’ve more than likely subverted the Copyright on the algorithm its self, so I imagine the original programmer could just sue you for that and show that anything you’ve created from it is a derivative work and demand royalties.

    2. I’m not sure what 3-d sculptures have to do with this…

      If you copy a program and use it to create something, the creation is yours. You do this all the time. The computer you’re using is running several copied algorithms created by other people, yet any works you create with your computer are considered your creation (or perhaps your boss’s creation if you’ve been employed to create things).

      The issue here really isn’t about copyright. It’s about Google being a near-monopoly as the dominant search engine. The legal case is being made preemptively, in the event of a government initiated antitrust case, that Google can’t be taken to task for the manner in which it orders its search results because it’s protected speech under the First Amendment.

  3. My take is that people have rights. Corporations, algorithms and books do not have rights. The newspaper editor has rights. The newspaper does not have rights. It is not a person. Copyright may be bought and sold, and have an independent life after the death of the creator, but 1st amendment rights cannot.

    Supreme Court verdict on rights of corporations is absurd ex facie.

    1. Except most of the major press outlets in this country are (and have long been) organized as corporations. If you said a corporation categorically can have no First Amendment rights whatsoever, freedom of the press would be getting exercised only by basement bloggers. It’s a corporation like the NY Times that has the money to send people to report on wars. We would lose something absolutely essential if corporations like the NYT, WaPo etc. didn’t enjoy the First Amendment right of freedom of the press.

      The problem with the Supreme Court’s decisions is that they are trending toward treating corporations as having *all* the same First Amendment rights as human beings. Even Hayek thought this was stupid (see his essay, “The Corporation in a Democratic Society: In Whose Interest Ought It To and Will It Be Run?”).

  4. TL:DR, but are Google’s search results already tweaked to the user’s apparent tastes? If two different people separately search the exact same term, they get very different Google search results based upon their “profile of interests.” This was just recently pointed out to me and I am not happy about it. This is the exact reason why I have avoided Bing all along: I CRAVE random search results, it feeds my brain’s need for exploration. But if my searches are tailored to my past search results, links I may have clicked, the types of things I have a tendency to click first, then my randomness is managed and I do. not. like.

    1. Make sure that when you use Google you aren’t logged into a Google account and you’ll get the particular flavor of non-randomness that you’re looking for.

      Yes, non-randomness.  Search results are not random – they are the result of a completely deterministic “popularity contest”.  What you’re getting is ultimately a ranking of the popularity of various search terms paired with pages on the web.  If you dial back the weight of your own preferences in comparison to the rest of the universe of Googlers, then you get a ranking based on the global popularity of connected pairs of term lists and pages.  If you dial up the weight of your own preferences, then you get a ranking based on your previous history of searches.  But it’s anything but random – by turning off the weights of your personal contributions what you’re getting is actually a guide towards the things that the majority of people using Google would be interested in when those terms are used as a query rather than “crazy random stuff”. 

  5. There is a tension between claiming that search results are protected speech because they are under a second order editorial control and that search results do not violate copyright because they are simply presenting what is found and not making an editorial selection. (AP and many media sources often argue the latter.) Obviously, the truth is somewhere in the middle.

    Also, isn’t it time to bring back search apps like Sherlock and Watson to get around the blinders of personalized search. Either that or per-window identity management. I’m constantly logging in and out of google and bing.

  6. Doctorow’s point on libel liability is well-taken.  I don’t think they can have it both ways–a “speaker” for antitrust purposes, but a mere “conduit” immunized by the DMCA and CDA when that status is convenient.

  7. Oh, and here’s a bit more analysis: 

    1) P. 1188 – 1201 of this article co-authored by UTex prof Oren Bracha and I addresses 1A issues: 2) Disclosure remedies don’t impinge on the search engines’ editorial judgment; rather, it’s about reporting on changes to ranking practices.  And they have already been applied (via Dodd-Frank) to rating agencies, which have 1st Amendment protections for their ratings (as opinions). Mark Patterson has a paper on that.3) Overall, I think the critical issue is going to be how the FTC and courts conceive of “editorial judgment.”  Here’s a critical part of the Volokh paper: “But the newspaper in Lorain Journal Co. was not excluding advertisements because of their content, in the exercise of some editorial judgment that its own editorial content was better than the proposed advertisements. Rather, it was excluding advertisements solely because the advertisers—whatever the content of their ads—were also advertising on a competing radio station.”My question for them would be: might this be said: Foundem [or other complainants] say Google was not excluding (or charging 100X more for reserve bids on) advertisements because of their content, in the exercise of some editorial judgment that its own editorial content was better than the proposed advertisements. Rather, they say it was excluding them because it wanted to promote its own subsidiaries, etc., instead.4) Of possible interest: a 2010 piece by Cade Metz, whose reporting was quite interesting: to an antitrust scholar at the University of Iowa, Google “opinion” talk could weaken its antitrust defense.

  8. First off, any search engine that applies the concept of stopwords is applying a filter that reflects a human’s decision on which words are to be disregarded in the content as “noise”.

    Search results ranking and media outlet ranking are similar. Instead of selecting a channel, you are requesting a particular type of content in a vague manner. If your selection provides a highly occurring term, like “shoes”,  a straightforward algorithm will bring up documents that have a high percentage of their characters arranged in the pattern “shoes” bounded on either side by a word break character (like a space, or punctuation).  A pure algorithm doesn’t know if you mean brake shoes, or the history of all manner of shoes, or shoes you wear on your feet, or songs and videos with “shoes” in the title. It’s just counting and adding.

    Google’s algorithm favors not only the meaning of shoes on your feet, but that you want to shop for them. So Google’s automatic editor believes that what most people mean when they ask simply “shoes” is “me want buy shoes for feet”.  Also, those results appear near instantaneously.

    Search for simply “biscuit”, however, and you get information about biscuits.  The automatic editor thinks you’re asking “what mean this word, ‘biscuit’?”. It also takes the results longer to appear, and all the while you are typing “biscuit” it is feverishly trying to complete your phrase, up until the last character believing your mosly likely choice to be “bisquick recipe”.  

    Plus various weight might be added to documents that are newer, or larger, or more often accessed, or all kinds of stuff.

    So, yes: the Google search is not a pure web data retrieval mechanism with a pure algorithm, even though it would be fun sometimes if it was, it would also be annoying a lot of the time, which of course is why they’ve made it to tweak the results, in all kinds of ways I would guess.

    At this point, as a side effect, it has become a biased information displayer with an agenda: an automatic editor.

Comments are closed.