Santa Clara University School of Law/Cyberspace Law Spring 1999
Professor Eric Goldman (firstname.lastname@example.org)
This document describes some of the issues I was looking for in answers to this year’s final exam. As usual, this document was not a checklist, but obviously the more issues addressed correctly, the higher the score. I gave 19 As, 42 Bs and 6 Cs.
I have now become a huge fan of the examinator. My average time reviewing and grading examinator-processed exams was less than 10 minutes, whereas some hand-written exams took over 30 minutes to read. Interestingly, while I did not do the math, I don’t think the examinator exams did better grade-wise (if anything, the opposite may have been true). However, for simple minimization of eye strain, the examinator was a clear winner in my book!
As usual, this was a real life situation. One of our clients got a letter from “GMM,” and this required our firm to more deeply examine the client’s practices than we had done prior to such time.
When pitched to me as a project, the partner described this as an easy case, requiring 15 minutes or so my time to give the client some quick support. Part of this misperception was predicated on the client’s belief that they were merely doing the same thing as “offline browsers.” Offline browsers work just like a normal browser, except that they can be programmed to grab http pages even when the user is not otherwise in a web session (i.e., off hours). While at the highest levels of abstraction, the software does perform the same functions as an offline browser, the technical implementation is not the same as an offline browser in ways that matter. Thus, after several hours of client interviews and self-investigation, I was able to piece together the technical operation in ways that made it clear we were dealing with a much more difficult animal.
I expected a good answer to address hot news, trespass, copyright and contract. Some of you discussed only 3 of the 4, which usually degraded your score some; some of you did not discuss trespass or copyright, and this definitely lowered your score. As I mentioned repeatedly in class, there was a lot of time pressure on this question, which made it impossible to reach every esoteric issue. However, some of you were very successful at sketching out the basic elements of the right causes of action and providing some decent factual analysis, and this usually was enough to place the answer into the A range.
While the entire hot news doctrine is pretty controversial, this factual situation appears to be as clearly a hot news violation as I think we can see:
(1) plaintiff generates or collects information at some cost or expense. It is fair to assume that GMM creates its own stories or licenses them at some expense.
(2) the value of the information is highly time-sensitive. News headlines should fall into this category. Note that the information was being refreshed every few minutes—another example of the time-sensitivity of the information.
(3) defendant’s use of information constitutes free-riding on plaintiff’s efforts. Ticketmaster asserted that Microsoft’s deep linking was a form of free riding. Whether we believe that or not, the real free riding here is the use of robots to collect published information as an alternative to developing the information oneself. Compare the INS v. AP situation, where representatives from the competing news agency transcribed stories from the wall and then sent them via telegraph. While the competing news agency incurred some cost to do so, the cost was negligible compared to the real cost incurred in developing the news.
(4) defendant’s use of information is in direct competition with plaintiff. While the services generally compete, the real competition was between the Flaschy service and the GMM email service, which are truly direct substitutes for each other.
(5) free-riding would reduce the incentive to create the product or service such that its existence or quality would be substantially threatened. This of course is impossible to determine on the facts that we have. Many of you argued that the Flaschy service created additional traffic for GMM and therefore was a net win for GMM. The client certainly believed this. However, this misses the point that headlines are a different “product” than the full text of the news, and GMM was able to sell banners both for the headlines aggregation pages and for the emails with its headlines, each of which would lose traffic as Flaschy’s service expanded. Thus, I believe this factor is likely to be met.
On balance, I believe that the hot news claim is very serious.
Some of you argued that the links in the ticker were a separate hot news violation. This is theoretically possible, but seems very remote, especially in comparison to the much stronger hot news claims coming from actually having the headlines in the ticker.
Some of you offered Shetland Times as an example of hot news problems. Shetland Times is a Scottish copyright case, not an American hot news case, and thus has limited precedential value.
As I mentioned in class, I generally believe the use of robots raises trespass issues. Robots almost always meet the 2 required elements of trespass:
(1) “intentionally using or intermeddling with the chattel in possession of another.” The electronic contacts of the robot meet this element.
(2) Plaintiff was dispossessed of the chattel; the chattel is impaired as to its condition, quality or value; Plaintiff was deprived of use of the chattel for a substantial period of time; or Plaintiff suffered bodily harm or harm is caused to some person or thing in which the plaintiff has a legally protected interest. Robots almost always deprive the plaintiff of the use of the chattel to some degree. Further, in this case, the harm caused by, for example, the hot news misappropriation or a copyright infringement would constitute harm to plaintiff’s legally protected interest.
CompuServe v. Cyber Promo implied 2 additional requirements to state a good trespass cause of action:
(3) Sufficient notice. The “terms” page may have constituted sufficient notice. Robot exclusion headers would likely constitute sufficient notice, but we don’t know if they were used. The cease and desist letter certainly constitutes sufficient notice.
(4) Technological self-help. In the CompuServe case, CompuServe had tried to block the IP address of the spam originator, only to see the spammer use fake IP addresses to send the spam. Here, we don’t know if GMM used technological self-help. Therefore, it’s uncertain if they met this element or not. It’s also uncertain if other courts are going to require this as an element (contrast Hotmail, which didn’t mention it).
Some of you asserted that the cease and desist letter constituted technological self help. In CompuServe, the court was talking about IP address blocking. Robot exclusion headers might also do the trick, or they may only be sufficient to convey notice. It’s hard to see how a paper letter would constitute self help.
Many of you talked about “implied licenses” for trespass. This is a confusing articulation. Some have argued that there is an implied consent to trespass if a site hooks up to the web without setting up technological barriers. Certainly this implied consent would apply to normal browsing unless expressly destroyed, but beyond that I’m not sure how far the implied consent goes. In some respects this is the flip side of the requirement for technological self-help.
On balance, GMM may have a strong case of trespass, depending on what technological self-help was actually used by GMM and how likely it is a future court would require this.
As discussed in class, there is a theoretical argument that unwanted linking, such as was occurring from the Flaschy client, could constitute trespass. This claim is a lot more dubious.
There are multiple actions that might constitute copyright infringement. The better answers kept these claims separate.
(1) Accessing the headlines pages. The robots made copies of the pages. While this creates a copy, this copying is indistinguishable from the copying made by any other browser. Thus, to find this an infringement, a court would have to find that all browsing was infringement. This is not a serious claim.
(2) Placing the headlines into the database. Unlike typical browsers, Flaschy’s robots distilled each accessed page down into a smaller unit. This is both a copy and possibly a derivative work. Flaschy’s retort is that the copied elements (headlines) are not copyrightable, and thus there was no copying of protected expression. While we usually believe that headlines are not copyrightable under Feist, GMM might assert a compilation copyright in the selection/arrangement/coordination of the headlines, which arguably Flaschy was infringing.
(3) Distributing the headlines. If the headlines were copyrightable, or if GMM had a protectable compilation copyright in the headlines, then Flaschy was probably infringing the copyright by reproducing, distributing and displaying the headlines through the client software.
(4) Linking to the full-text articles. As discussed in class, linking is not a direct copyright infringement. Some pundits have asserted that linking is contributory infringement, but this is predicated on the users making an infringing copy, which we continue to believe cannot be the state of the law. This is not a serious claim.
Some of you argued there may be vicarious copyright infringement for linking, an argument I’ve not heard before. Without reaching the issue of a linking site’s right and ability to control the infringement (I think that is very confusing), it’s hard to construct an argument that a non-framing site derives direct financial benefit from the infringing activity (i.e., linking to a remote site).
Thus, the most serious claim is that Flaschy is infringing GMM’s copyright by making a derivative work of GMM’s pages and then distributing that derivative work using the software. Assuming that GMM has protectable interests here, then Flaschy will assert that it has an implied license to do so or its behavior is fair use.
While the browsing-style infringements might well be subject to an implied license argument, the specific behavior of distilling a page down and distributing the distilled content is much less likely to be subject to an implied license. There is also a decent argument that the “terms” page destroys the implied license.
A number of you argued that the copyright notice at the bottom of the page destroyed any implied license theory. I don’t think this is right. Asserting that one has a copyright interest is much different from limiting by notice or contract how the interest may be used. Otherwise, to the extent that browsing is subject to an implied license, the copyright notice would cut that theory off.
As for fair use:
- Flaschy’s use is commercial
- GMM’s information is published, and Flaschy is taking the news-like portions of the content
- Flaschy is taking 100% of the headlines. Some of you argued that it was taking none of the full text of the articles, or that the headlines were only a part of the overall content. Again, this fails to acknowledge that the headlines were their own separate product.
- Flaschy’s use has a detrimental effect on the market for GMM’s headlines product, as discussed in the Hot News discussion.
On balance, Flaschy is probably going to lose fair use, if it gets that far. However, overall the copyright claim is pretty weak, since Flaschy’s copying is limited to areas where GMM has thin copyright protection in any event.
There are 2 other copyright claims to consider:
(5) Flaschy could be subject to a criminal copyright claim. While Flaschy should meet the financial gain test and the dollar test, the issue is whether Flaschy’s infringement is willful. Certainly Flaschy’s conduct is willful, but Flaschy was trying not to take copyrighted material in its conduct. Thus, the criminal copyright claim is weak.
(6) DMCA Copyright Management Information claim (17 USC 1202). We may very well have a 1202 violation here, since Flaschy does not propagate any CMI as it disseminates the GMM headlines. Flaschy’s assertion (if true) that it ends up distributing only non-copyrighted elements of GMM’s content may be a good defense to the 1202 claim, but I did not research this.
Many of you discussed 17 USC 512 (the DMCA safe harbors) in this question. The safe harbors offer no help here. 512(a) applies to ISPs. 512(b) applies to caching, which arguably could apply here except that the safe harbor applies to content initially being moved at the request of a user. 512(c) applies to material stored at the direction of a user. 512(d) applies to linking to infringing content. Thus, none of these safe harbors apply to a direct infringer who is engaging, at their own initiative, in infringing activity.
Some of you asserted that the client software could lead to a 1201 violation. The software itself was not intended to circumvent any technological protection devices, so this seems very remote.
Some of you confused the changes in remedies for copyright notices and copyright registrations. Copyright notices merely cut off the partial defense of innocent infringement, which affects statutory damage calculations. Copyright registrations can create the opportunity to get statutory damages and attorneys fees. There was no evidence of copyright registration here.
Breach of Contract
Flaschy is probably breaching the “terms” if these constitute a contract. However, since the terms were not displayed as a non-leaky clickthrough, it is likely they will not be treated as a contract. There is a secondary approach that might be relevant—that Flaschy clicked through the pages anyway, even though it was not mandatory, and thus manifested assent to them anyway—but this would require a number of factual assumptions and would be difficult to prove. Some of you discussed whether the robot could create a contract, although there was no evidence that the robots were sent to the terms pages.
Other Possible Claims
- Trademark Infringement. There may be an implied sponsorship between the parties. Also, there could be a reverse passing off style claim, where Flaschy is falsely taking credit as the source/origin of the product when it really is not the source/origin.
- Unfair competition. This could be supported by a number of Flaschy’s actions.
A few of you created a separate deep linking cause of action. I don’t believe this exists, but rather deep linking is subsumed within the other claims described above.
Some of you argued that there may be an intentional interference with prospective economic advantage. While there certainly could be a negative effect on GMM’s ability to obtain advertising, nothing in the facts supported this claim. This type of claim, on these facts, probably devolves into an unfair competition or misappropriation claim.
Changes to Flaschy’s Practices
- Obtain licenses. The most obvious solution is to obtain a license to the headlines. In our case, we prepared a short form of “consent letter” which would permit Flaschy to get the headlines and redistribute them. The letter was really short with the objective of having it avoid being run through the legal department. Interestingly, if Flaschy grew large enough to be a significant distribution base, over time the information publishers might be willing to pay Flaschy for visibility to Flaschy’s user base.
- Stop sending robots to GMM without consent. This alleviates the immediate danger of a trespass claim from GMM based on the cease and desist letter.
- Honor robot exclusion headers and do not forge the IP address of the robots. These diminish the trespass claim.
- Use humans to collect and rewrite the headlines. This may reduce the trespass claim (although arguably the human use is no different or less imposing than the robot accesses), although note that it does not diminish the hot news claim at all (see INS v. AP).
- Obtain the headlines through GMM’s email service as opposed to from the website. This may diminish the trespass claim, although it does nothing with respect to the hot news claim, and it may exacerbate the contract claim if GMM uses a mandatory clickthrough as part of the sign-up process for the email service.
- Time-delay the redistribution of headlines. This would diminish the hot news claim, although it makes the service less valuable.
- Incorporate publishers’ copyright management information into the redistributed content. This mitigates the 1202 claim.
- Remove any indicia of sponsorship by the publishers. This was an issue in the real life situation, since the clients’ website contained some very ambiguous statements implying sponsorship which we needed to change.
- Instruct employees not to review the “terms” page. This diminishes the possibly esoteric claims regarding contract formation.
This was such an easy question! Or so I thought.
This type of issue comes up ALL THE TIME in my practice. Two parties are trying to do some form of content exchange, often without any monetary exchanges. And then one party freaks out and wants an indemnity for every possible cause of action that has ever been asserted in connection with the Internet, whether it actually could arise in the particular situation or not. This drives me bonkers. However, through a clear understanding of the application of the statutes, it becomes a little easier to quantify the risks for clients—especially where the risk is really small! In these cases, you can add a lot of value to the negotiations by cutting through all of the indemnity hoopla and accepting what appears to be a lot of risk for your clients based on a thorough grounding of how the statutes provide insulation for them.
The threshold issue is whether a website can be deemed an interactive computer service under 47 USC 230. Our working theory is that a website is an ICS, in which case SE is eligible to get the benefit of the 230 safe harbors. The fact that SE organizes the page layout should not have the effect of causing Routers’ information to be “provided by” SE (see Blumenthal). Routers undoubtedly is an information content provider under that statute.
(1) Defamation. 230 acts as a complete bar to SE’s liability for defamatory content from Routers. SE does not need an indemnity for defamatory content.
(2) Inaccurate Information. The question is whether SE would be deemed the “speaker or publisher” of Routers’ inaccurate content. My working theory is that inaccurate information would be deemed a publisher/speaker claim and thus would be preempted by 230. Even if it was not, theories such as those applied in the 9th Circuit mushroom case or in Daniel v. Dow Jones case (requiring a “special relationship” to establish a claim for inaccurate information) make it unlikely that SE would be liable even if 230 did not apply. SE can live without an indemnity for inaccurate information.
(3) Right of publicity/privacy. The question is whether this looks more like a “publisher/speaker” claim (covered under 230(c)) or an intellectual property claim (not covered, per 230(d)(2)). I don’t know the answer to this. My instincts are that this would be treated as an IP claim and thus not be covered under 230(c). Thus, although the client might be willing to risk that the claim is preempted by 230(c), we probably want an indemnity for this.
(4) Hot News. Same issue as the right of publicity/privacy claims, except I’m more confident that 230(d)(2) would apply to the hot news claims. Thus, we should push for an indemnity on this one.
(5) Obscenity/Harmful to Minors. 230(d)(1) says that federal criminal law is not covered by 230(c). Thus, SE theoretically could be liable for federal criminal claims for obscene or harmful to minor material. However, at the moment, there is no such thing as “harmful to minors” in the Internet context, as the federal law was struck down as unconstitutional in ACLU v. Reno II. Further, 230(c) should knock out all state law claims related to this topic (see, e.g., Doe v. AOL, applying 230(c) to certain state law claims involving child pornography). Thus, all we may need to worry about is obtaining an indemnity for federal criminal obscenity claims. With respect to these, we know that Smith v. CA will require at least some scienter before SE could be liable for these claims anyway. Further, since the content feed probably does not include pornography as an expected part of the data feed, the likelihood of any content rising to the level of obscenity seems very small. Finally, it is probably against public policy for Routers to indemnify SE for criminal violations (and jail time, the primary remedy in criminal obscenity claims, is not “indemnifiable” at all), so the indemnity probably won’t work anyway. Thus, SE is not taking on very much risk by skipping this indemnity.
Some of you argued that the “know/should have known” standard, if applied to these claims, makes the indemnity unnecessary because SE will not be liable until it receives some sort of notice, at which point it can manage the risk by taking the allegedly violative content down. Generally, I think this is right. Query, however, to what extent these claims are strict liability claims. For example, if the products liability claim is a strict liability claim, it’s not clear that the scienter requirement will be interposed with respect to the distributor (it should be under the First Amendment, but who knows?). To the extent that 230(c) speaks, it does much better for us than the default rule.
Some of you argued that 230’s protections depended on the amount of editing exercised by the ICS. This is not true. However, an argument can be made that some level of editing may cut off the argument that the information was “provided by” an ICP, although we don’t know the scope of this.
Some of you tried to argue that 230 would cut off only some forms of derivative liability but then talked about other forms of derivative liability. If 230 applies, it cuts off all forms of derivative liability.
A number of you discussed the DMCA in response to this question. The DMCA was irrelevant to this question. The DMCA sets up some safe harbors for copyright infringement, but I didn’t ask you about copyright infringement in this question. The DMCA offers no precedential value for the hot news analysis, for example.