Joe JAMES, et al., Plaintiffs-Appellants, v. MEOW MEDIA, INC., et al., Defendants-Appellees, 300 F.3d 683

OPINION BOGGS, Circuit Judge. On December 1, 1997, Michael Carneal walked into the lobby of Heath High School in Paducah, Kentucky, and shot several of his fellow students, killing three and wounding many others. The parents and estate administrators of Carneal’s victims — Jessica James, Kayce Steger, and Nicole Hadley — (hereinafter collectively referred to as “James”) appeal the judgment of the district court dismissing, for failing to state claims on which relief could be granted, their actions against several video game, movie production, and internet content-provider firms.

According to James’s complaint, Carneal regularly played video games, watched movies, and viewed internet sites produced by the defendant firms. These activities, James argues, “desensitized” Carneal to violence and “caused” him to kill the students of Heath High School. James claims that the distribution of this material to impressionable youth like Carneal constitutes actionable negligence under Kentucky law, entitling James to recover wrongful death damages from the distributing firms.

Moreover, James contends that the defendant firms purveyed defective “products,” namely the content of video games, movies, and internet sites, triggering strict product liability under Kentucky law. The defendant firms argue that they owe no duty to protect third parties from how players, watchers or viewers process the ideas and images presented in their video games, movies, and internet sites. Specifically, the defendants contend that Carneal’s actions were not sufficiently foreseeable to trigger the defendants’ liability.

Even if they were to owe such a duty to protect third parties from the consumers of their ideas and images, the defendants argue that Carneal’s independent decision to kill his fellow students constitutes a superseding cause of the claimed damages and defeats the proximate cause element of James’s prima facie case. The defendants further contend that tort liability for the non-defamatory ideas and images communicated in their respective media would raise significant First Amendment questions that ought to be avoided.

Finally, the defendants note that James’s theory of product liability is flawed as they have not distributed “products” under Kentucky law. For the reasons set forth below, we affirm the district court’s dismissal of James’s actions. I On December 1, 1997, Michael Carneal brought a .22-caliber pistol and five shotguns into the lobby of Heath High School in Paducah, Kentucky. At the time, Car-neal was a 14-year-old freshman student at the school. Upon his arrival, Carneal began shooting into a crowd of students, wounding five. His shots killed an additional three: Jessica James, Kayce Steger, and Nicole Hadley.

Carneal was arrested and convicted of murdering James, Steger, and Hadley. Subsequent investigations revealed that Carneal regularly played “Doom,” “Quake,” “Castle Wolfenstein,” “Redneck Rampage,” “Nightmare Creatures,” “Mech Warrior,” “Resident Evil,” and “Final Fantasy,” which are interactive computer games that, in various ways, all involve the player shooting virtual opponents. Car-neal also possessed a video tape containing the movie, “The Basketball Diaries,” in a few minutes of which the high-school-student protagonist dreams of killing his teacher and several of his fellow classmates.

Investigators examined Carneal’s computer and discovered that he had visited “www.persiankitty.com,” which appears to catalogue and link to sites with sexually-suggestive material. It also appeared that through “www.adultkey.com,” www. adult-key, com, a site operated by Network Authentication Systems and designed to restrict access to certain websites to viewers over certain ages, Carneal was granted age verification sufficient to visit many other pornographic sites.

The parents, who also are the estate administrators, of James, Steger, and Had-ley filed this action in the United States District Court for the Western District of Kentucky. James’s complaint named as defendants the companies that produce or maintain the above-mentioned movie, video games, and internet sites. James stated essentially three causes of action against the defendants. First, James alleged that the defendants had been negligent in that they either knew or should have known that the distribution of their material to Carneal and other young people created an unreasonable risk of harm to others.

James alleged that exposure to the defendants’ material made young people insensitive to violence and more likely to commit violent acts. But for Carneal’s steady diet of the defendants’ material, James contended, Carneal would not have committed his violent acts. Second, James asserted that the video game cartridges, movie cassettes, and internet transmissions that the defendants produced and distributed were “products” for purposes of Kentucky product liability law. According to James, the violent features of the movie, games, and internet sites were product defects.

The defendants, as producers and distributors of the “products,” are strictly liable under Kentucky law for damages caused by such product defects. The third claim that James raised was that some of the defendants’ conduct had violated the Racketeer Influenced and Corrupt Organizations Act, 18 U.S.C. § 1961 et seq. According to James, the defendant firms that maintained internet sites engaged in a pattern of racketeering activity by distributing obscene material to minors. The defendants moved to dismiss all of James’s actions for failing to state any claim on which relief could be granted.

See Fed.R.Civ.P. 12(b)(6). The district court granted the defendants’ motion. The district court held that Carneal’s actions were not sufficiently foreseeable to impose a duty of reasonable care on the defendants with regard to Carneal’s victims. Alternatively, the district court held that even if such a duty existed, Carneal’s actions constituted a superseding cause of the victims’ injuries, defeating the element of proximate causation notwithstanding the defendants’ negligence.

With regard to James’s second cause of action, the district court determined that the “thoughts, ideas and images” purveyed by the defendants’ movie, video games, and internet sites were not “products” for purposes of Kentucky law and therefore the defendants could not be held strictly liable for any alleged defects. Finally, the district court held that James had not stated a viable RICO claim against the defendant internet firms.

Among other reasons, the district court explained that James had alleged RICO predicate acts, but had failed to identify an organization that had been corrupted, and had failed to show any injury to “business or property” as required for standing under RICO. James now appeals the district court dismissal of his negligence and product liability claims. We can discern no argument in James’s brief assigning error to the district court’s dismissal of his RICO claim, and therefore consider such an argument waived. Accordingly, this appeal raises questions of only Kentucky law.

II In this case, we review a district court’s dismissal of a case for failure to state a claim on which relief may be granted. Fed R. Civ. P. 12(b)(6). In deciding a Rule 12(b)(6) motion to dismiss, the district court is required to assume that all allegations in the complaint are true and to decide whether those allegations provide a basis for legal relief. See Mayer v.

Mylod, 988 F.2d 635, 638 (6th Cir.1993). “A complaint should not be dismissed for failure to state a claim unless it appears beyond doubt that the plaintiff can prove no set of facts in support of his claim which would entitle him to relief.” Conley v. Gibson, 355 U.S. 41, 45-46, 78 S.Ct. 99, 2 L.Ed.2d 80 (1957). This court reviews de novo a district court’s granting of such a motion to dismiss. Helfrich v. PNC Bank, Kentucky, Inc., 267 F.3d 477, 480-81 (6th Cir.2001). The legal questions under review in this case are questions of Kentucky law.

When faced with the resolution of these state law issues, our inquiry is in part hypothetical. Specifically, we are charged to decide these questions of Kentucky state law as Kentucky state courts would. See Miles v. Kohli & Kaliher Assocs., 917 F.2d 235, 241 (6th Cir.1990). Although we can find no decision of Kentucky courts that decides precisely the issues presented in this case, we are confident that Kentucky appellate courts would uphold the dismissal of James’s action, based on their existing decisions.

We elaborate on our conclusion below, considering first James’s negligence action and second his products liability action. Ill James contends that the defendants in this case acted negligently, perhaps in producing, but at least in distributing to young people, their materials. It was this negligence, according to James, that caused Carneal to undertake his violent actions and that thereby caused the deaths of the plaintiffs’ daughters.

In order to establish an actionable tort under Kentucky law, the plaintiff must establish that the defendant owed a duty of care to the plaintiff, that the defendant breached that duty of care, and that the defendant’s breach was the proximate cause of the plaintiffs damages. A. The Existence of a Duty of Care The district court held that James’s allegations, even if assumed to be true, failed to establish the first element of the prima facie case. Specifically, the district court determined that the defendants were under no duty to protect James, Steger, and Hadley from Carneal’s actions.

James argues that the district court erred as a matter of Kentucky law in this regard. The “existence of a duty of care” as an element of a tort cause of action is of relatively recent vintage. At early English common law, the existence of a duty of care was not considered as an element of an actionable tort. Then, tort law attached strict liability for any damages that resulted from “wrongful acts.” Prosser on Torts § 53. The requirement that the plaintiff establish that he was owed a specific duty of care by the defendant came with the advent of negligence in place of strict liability. See generally Percy H.

Winfield, Duty in Tortious Negligence, 34 Colum. L.Rev. 41 (1934). As a separate basis of liability, negligence was simply too broad, according to later English courts. Accordingly, the concept of limited duty disciplined the concept of negligence, requiring the plaintiff to establish a definite legal obligation. At first, the limited duty of care concerned the persons to whom a defendant owed an obligation of care. See, e.g., Shore v.

Town of Stonington, 187 Conn. 147, 444 A.2d 1379 (1982) (holding that a police officer owed no duty to protect a citizen who was killed by a drunk driver that the officer had stopped, but released). The dilemma was that it became nearly impossible categorically to demarcate the universe of people to whom a person would owe a duty to abstain from negligence. With regard to the existence of duty between specific persons, the answer always was dependent on the circumstances. Kentucky courts resolved this dilemma by establishing a “universal duty of care.” Grayson Fraternal Order of Eagles v.

Claywell, 736 S.W.2d 328 (Ky.1987); M & T Chems., Inc. v. Westrick, 525 S.W.2d 740 (Ky.1974); Greyhound Corp. v. White, 323 S.W.2d 578 (Ky.1958). Under the universal duty of care, “every person owes a duty to every other person to exercise ordinary care in his activities to prevent foreseeable injury.” Grayson, 736 S.W.2d at 332. James argues that Kentucky’s universal duty of care would foreclose the defendants’ argument that they had no duty to protect James, Steger, and Hadley.

It is clear to us, however, that Kentucky’s rule of a universal duty of care simply recognizes the obvious in determining the existence of a duty of care. Of course, there are conceivable circumstances under which there would be legally enforceable duties to act without negligence between any two people. What Kentucky courts have recognized through the “universal duty of care” rule is that the existence of a duty of care is circumstantially hmited: the duty is to exercise ordinary care to prevent foreseeable harm.

Thus, Kentucky courts have held that the determination of whether a duty of care exists is whether the harm to the plaintiff resulting from the defendant’s negligence was “foreseeable.” Foreseeability is an often invoked, but not terribly well defined, concept in the common law of tort. Some common law tort regimes use foreseeability as the standard for determining proximate causation. See, e.g., Fu v. State, 263 Neb. 848, 643 N.W.2d 659, 669 (2002); Haliw v. City of Sterling Heights, 464 Mich. 297, 627 N.W.2d 581, 588 (2001).

Compare Bodkin v. 5401 S.P., Inc., 329 Ill.App.3d 620, 263 Ill.Dec. 434, 768 N.E.2d 194, 202 (2002) (attempting to identify different forms of foreseeability for the proximate cause and the existence of a duty inquiries). Kentucky’s particular use of foreseeability in the duty inquiry finds its roots in perhaps the most famous application of the foreseeability principle. In Palsgraf v. Long Island Railroad Co., 248 N.Y. 339, 162 N.E. 99 (1928), then-judge Cardozo determined that the defendant’s duty is to avoid “risks reasonably to be perceived.” Id. at 100.

As every former law student remembers, the plaintiff in Palsgraf had bought a train ticket to travel to the Rockaways for the afternoon and was waiting on the station platform. Two men were rushing to catch a departing train, one easily hopped on board and the other struggled to pull himself onto the rear car. A conductor on the train pulled him on board, but in the process, the struggling passenger dropped onto the rails the brown-paper package he was carrying. The package was full of fireworks and exploded.

The explosion overturned a large set of scales on the platform, which struck Palsgraf, and Palsgraf sued the railroad for her injuries. Cardozo determined that the railroad simply did not owe a duty to Palsgraf to protect against the injury that she suffered. For Cardozo, the harm that Pals-graf suffered was not sufficiently probable that the railroad employees could have been expected to anticipate it occurring from their actions.

Cardozo’s reasoning, although implying that Palsgraf was the unforeseeable plaintiff, rested on the improbability of the harm that she suffered arising from the defendant’s particular actions. Id. at 101. For Cardozo too, the existence of a duty of care was a creature of circumstance. Cardozo’s opinion in Palsgraf, while cited as the cornerstone of the American doctrine of a limited duty of care, has been criticized for its conclusory reasoning regarding whether Palsgrafs harm really was sufficiently unforeseeable. See William L. Prosser, Palsgraf Revisited, 52 Mich. L.Rev. 1, 7-9 (1953).

Such concluso-ry reasoning has been endemic in the jurisprudence of determining duty by assessing foreseeability. Courts often end up merely listing factual reasons why a particular harm, although having materialized, would have appeared particularly unlikely in advance and then simply asserting that the harm was too unlikely to be foreseeable and to create a duty to exercise due care in protecting against it. What has not emerged is any clear standard regarding what makes a projected harm too improbable to be foreseeable. Kentucky courts have struggled with the formless nature of this inquiry.

At bottom, Kentucky courts have conceded, deciding the existence of a duty of care is “essentially a policy determination.” Mullins v. Commonwealth Life Ins. Co., 839 S.W.2d 245, 248 (Ky.1992); Sheehan v. United Servs. Auto. Ass’n, 913 S.W.2d 4, 6 (Ky.Ct.App.1996). Rather than the sophisticated weighing of probabilities, the content of that policy determination “is but a conclusion of whether a plaintiffs interests are entitled to legal protection against the defendant’s conduct.” Sheehan, 913 S.W.2d at 6.

The parties in this case have argued at length over whether the foreseeability inquiry required to determine the existence of a duty of care is a pure question of law for the court or a question of fact that should generally be submitted to a jury. Under Kentucky law, it is clear that the existence of a duty of care to the plaintiff, and its underlying foreseeability inquiry, is a pure question of law for the court. Mullins, 839 S.W.2d at 248 (“The question of duty presents an issue of law”); Sheehan, 913 S.W.2d at 6.

See also 57A Am.Jur.2d Negligence § 20 (listing other jurisdictions in which the duty inquiry is a question of law). The allocation of responsibility for determining this question to the courts, rather than to juries, reveals that the duty inquiry contains an important role for considering the policy consequences of imposing liability on a certain class of situations. Essentially, the foreseeability inquiry requires courts to determine questions inexorably tied to the ultimate question of whether the defendant was negligent.

After all, the probability of the harm is a significant factor in traditional assessment of negligence. By placing the foreseeability analysis in the hands of courts, the existence of duty element of the prima facie case serves as a gatekeeper for the otherwise extremely broad concept of negligence. Thus, we are called, as best we can, to implement Kentucky’s duty of care analysis in this case. Our inquiry is whether the deaths of James, Steger, and Hadley were the reasonably foreseeable result of the defendants’ creation and distribution of their games, movie, and internet sites.

Whether an event was reasonably foreseeable is not for us to determine with the assistance of hindsight. Mitchell v. Hadl, 816 S.W.2d 183, 186 (Ky.1991). The mere fact that the risk may have materialized does little to resolve the foreseeability question. Kentucky courts, in resolving foreseeability questions, have consistently inquired into the relative likelihood of the injuries produced. Of particular interest in this case are cases in which plaintiffs have sought to hold defendants liable for the actions of third parties, allegedly enabled or encouraged by the defendants.

A line of cases in this vein concerns dram shops. In Grayson, the Kentucky Supreme Court held that an automobile accident injuring third parties was a reasonably foreseeable result of the negligent act of serving alcohol to an intoxicated individual. 736 S.W.2d at 332. In contrast, the court later held that an intoxicated patron fighting with and shooting a fellow customer was simply not a foreseeable result of continuing to serve the patron alcohol. Isaacs v. Smith, 5 S.W.3d 500, 502 (Ky. 1999).

According to the court, the violent fighting and shooting was so much less likely a result from the serving of alcohol than the negligent operation of a motor vehicle that it was not reasonably foreseeable. Ibid. Accordingly, the court held that dram shops, although negligent in serving alcohol to the obviously intoxicated, do not have a duty to protect third parties from the intentional violent acts of their intoxicated patrons. Ibid. Intentional violence is less likely to result from intoxication than negligent operation of a motor vehicle.

Yet, the Kentucky Supreme Court never makes clear how unlikely is too unlikely for a particular type of harm to be unforeseeable. The cases do not create a principle, portable to the context of this case, for evaluating the probability of harm. This court has encountered this foreseeability inquiry under Kentucky law before in a situation similar to this case. In Watters v. TSR, Inc., 904 F.2d 378 (6th Cir.1990), the mother of a suicide victim sued TSR for manufacturing the game “Dungeons and Dragons.” The suicide victim regularly played the game.

The mother contended that the game’s violent content “desensitized” the victim to violence and caused him to undertake the violent act of shooting himself in the head. We held that the boy’s suicide was simply not a reasonably foreseeable result of producing the game, notwithstanding its violent content. To have held otherwise would have been “to stretch the concepts of foreseeability and ordinary care to lengths that would deprive them of all normal meaning.” Id. at 381. Foreseeability, however, is a slippery concept. Indeed, it could be argued that we ourselves confused it with some concept of negligence.

We noted in Wat-ters: “The defendant cannot be faulted, obviously, for putting its game on the market without attempting to ascertain the mental condition of each and every prospective player.” Ibid. We almost appeared to say that the costs of acquiring such knowledge would so outweigh the social benefits that it would not be negligent to abstain from such an investigation. We can put the foreseeability point a little more precisely, however. It appears simply impossible to predict that these games, movie, and internet sites (alone, or in what combinations) would incite a young person to violence.

Carneal’s reaction to the games and movies at issue here, assuming that his violent actions were such a reaction, was simply too idiosyncratic to expect the defendants to have anticipated it. We find that it is simply too far a leap from shooting characters on a video screen (an activity undertaken by millions) to shooting people in a classroom (an activity undertaken by a handful, at most) for Car-neal’s actions to have been reasonably foreseeable to the manufacturers of the media that Carneal played and viewed. At first glance, our conclusion also appears to be little more than an assertion.

Mental health experts could quite plausibly opine about the manner in which violent movies and video games affect viewer behavior. We need not stretch to imagine some mixture of impressionability and emotional instability that might unnaturally react with the violent content of the “Basketball Diaries” or “Doom.” Of course, Carneal’s reaction was not a normal reaction. Indeed, Carneal is not a normal person, but it is not utter craziness to predict that someone like Carneal is out there.

We return, however, to the Kentucky court’s observation that the existence of a duty of care reflects a judicial policy judgment at bottom. From the Kentucky cases on foreseeability, we can discern two relevant policies that counsel against finding that Carneal’s violent actions were the reasonably foreseeable result of defendants’ distribution of games, movies, and internet material. 1.

The Duty to Protect Against Intentional Criminal Actions First, courts have held that, except under extraordinary circumstances, individuals are generally entitled to assume that third parties will not commit intentional criminal acts. See Restatement (Second) of Torts § 302B, cmt. d (“Normally the actor has much less reason to anticipate intentional misconduct than he has to anticipate negligence. In the ordinary case he may reasonably proceed upon the assumption that others will not interfere in a manner intended to cause harm to anyone.

This is true particularly where the intentional conduct is a crime, since under ordinary circumstances it may reasonably be assumed that no one will violate the criminal law.”). See also Gaines-Tabb v. ICI Explosives USA, Inc., 995 F.Supp. 1304 (D.Okla.1996) (holding that fertilizer and blasting cap manufacturers were not liable for Murrah Federal Building bombing, as they were entitled to believe that third parties would not engage in intentional criminal conduct); Henry v. Merck & Co., 877 F.2d 1489, 1493 (10th Cir.1989). The reasons behind this general rule are simple enough.

The first reason is a probabilistic judgment that foreseeability analysis requires.. Individuals generally are significantly deterred from undertaking intentional criminal conduct given the sanctions that can follow. The threatened sanctions make the third-party intentional criminal conduct sufficiently less likely that, under normal circumstances, we do not require the putative tort defendant to anticipate it. Indeed, this statistical observation explains the distinction drawn by Kentucky courts in the dram shop liability cases. The second reason is structural.

The system of criminal liability has concentrated responsibility for an intentional criminal act in the primary actor, his accomplices, and his co-conspirators. By imposing liability on those who did not endeavor to accomplish the intentional criminal undertaking, tort liability would diminish the responsibility placed on the criminal defendant. The normative message of tort law in these situations would be that the defendant is not entirely responsible for his intentional criminal act.

Does this case involve the extraordinary circumstances under which we would require the defendants to anticipate a third party’s intentional criminal act? Kentucky courts have found such circumstances when the tort defendant had previously developed “a special relationship” with the victim of a third-party intentional criminal act. See, e.g., Fryman v. Harrison, 896 S.W.2d 908, 910 (Ky.1995) (requiring that the victim be in state custody in order to trigger governmental duty to protect her from third-party violence as a matter of Kentucky tort law). Cf. DeShaney v. Winnebago County Dep’t of Soc.

Serv., 489 U.S. 189, 109 S.Ct. 998, 103 L.Ed.2d 249 (1989) (articulating the special relationship test to trigger a governmental duty to protect persons from private violence under the Fourteenth Amendment). This duty to protect can be triggered by placing the putative plaintiff in custody or by taking other affirmative steps that disable the plaintiff from protecting himself against third-party intentional criminal acts. Of course, a special relationship can be created by a contract between the plaintiff and the defendant.

Finally, some states have imposed a duty to protect others from third-party intentional criminal acts on members of discrete professions who become aware of the third-party’s intention to engage in criminal conduct against a specific person. See Tarasoff v. Board of Regents of the Univ. of Cal., 17 Cal.3d 425, 131 Cal.Rptr. 14, 551 P.2d 334 (1976); Evans v. Morehead Clinic, 749 S.W.2d 696 (Ky.Ct.App.1988). We can find nothing close to a “special relationship” in this case.

The defendants did not even know James, Steger, and Hadley prior to Carneal’s actions, much less take any affirmative steps that disabled them from protecting themselves. Courts have held, under extremely limited circumstances, that individuals, notwithstanding their relationship with the victims of third-party violence, can be liable when their affirmative actions “create a high degree of risk of [the third party’s] intentional misconduct.” Restatement of Torts (Second) § 302B, cmt. e.H.

Generally, such circumstances are limited to cases in which the defendant has given a young child access to ultra-hazardous materials such as blasting caps, Vills v. City of Cloquet, 119 Minn. 277, 138 N.W. 33 (1912), or firearms. Spivey v. Sheeler, 514 S.W.2d 667 (Ky.1974). Even in those cases, courts have relied on the third party’s severely diminished capacity to handle the ultra-hazardous materials.

With older third parties, courts have found liability only where defendants have vested a particular person, under circumstances that made his nefarious plans clear, with the tools that he then quickly used to commit the criminal act. See Meers v. McDowell, 110 Ky. 926, 62 S.W. 1013 (1901). Arguably, the defendants’ games, movie, and internet sites gave Carneal the ideas and emotions, the “psychological tools,” to commit three murders. However, this case lacks such crucial features of our jurisprudence in this area.

First, the defendants in this case had no idea Carneal even existed, much less the particular idiosyneracies of Carneal that made their products particularly dangerous in his hands. In every case that this court has discovered in which defendants have been held liable for negligently creating an unreasonably high risk of third-party criminal conduct, the defendants have been specifically aware of the peculiar tendency of a particular person to commit a criminal act with the defendants’ materials.

Second, no court has ever held that ideas and images can constitute the tools for a criminal act under this narrow exception. Beyond their intangibility, such ideas and images are at least one step removed from the implements that can be used in the criminal act itself. In the cases supporting this exception, the item that the defendant has given to the third-party criminal actor has been the direct instrument of harm. 2.

First Amendment Problems Moreover, we are loath to hold that ideas and images can constitute the tools for a criminal act under this exception, or even to attach tort liability to the dissemination of ideas. We agree with the district court that attaching tort liability to the effect that such ideas have on a criminal actor would raise significant constitutional problems under the First Amendment that ought to be avoided. See NLRB v.

Catholic Bishop of Chicago, 440 U.S. 490, 500, 99 S.Ct. 1313, 59 L.Ed.2d 533 (1979) (suggesting that non-constitutional questions of law ought to be construed, where possible, to avoid significant constitutional problems). Although the plaintiffs’ contentions in this case do not concern the absolute proscription of the defendants’ conduct,- courts have made clear that attaching tort liability to protected speech can violate the First Amendment. See New York Times v.

Sullivan, 376 U.S. 254, 265, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964) (“Although this is a civil lawsuit between private parties, the Alabama courts have applied a state rule of law which petitioners claim to impose invalid restrictions on their constitutional freedoms of speech and press. It matters not that that law has been applied in a civil action and that it is common law only, though supplemented by statute.”). The first inquiry is whether the defendants’ activity constitutes protected speech under the First Amendment.

One thing is perfectly clear to this court: the plaintiffs’ argument does not seek to attach liability to the cassettes and cartridges distributed by the defendants, but the ideas and images communicated to Carneal by those products. Although the defendants’ products may be a mixture of expressive and inert content, the plaintiffs’ theory of liability isolates the expressive content of the defendants’ products. Expression, to be constitutionally protected, need not constitute the reasoned discussion of the public affairs, but may also be for purposes of entertainment. See Time, Inc. v.

Hill, 385 U.S. 374, 388, 87 S.Ct. 534, 17 L.Ed.2d 456 (1967). Clearly, the various media distributed in this case fall along a spectrum of expressive content. It is long settled that movies can constitute protected speech. Joseph Burstyn, Inc. v. Wilson, 343 U.S. 495, 501, 72 S.Ct. 777, 96 L.Ed. 1098 (1952). Of more recent, but no less definitive, resolution is that internet sites are similarly entitled to protection. See Reno v. American Civil Liberties Union, 521 U.S. 844, 117 S.Ct. 2329, 138 L.Ed.2d 874 (1997) (striking down the Communications Decency Act regulating indecent material on the internet).

The constitutional status of video games has been less litigated in federal courts. Yet most federal courts to consider the issue have found video games to be constitutionally protected. American Amusement Mach. Ass’n v. Kendrick, 244 F.3d 572 (7th Cir.2001) (Posner, J.); Wilson v. Midway Games, Inc., 198 F.Supp.2d 167 (D.Conn.2002); Sanders v. Acclaim Entm’t, Inc., 188 F.Supp.2d 1264 (D.Colo. 2002). But see Interactive Digital Software Ass’n v. St. Louis County, Mo., 200 F.Supp.2d 1126 (E.D.Mo.2002) (holding that video games were not protected expression under the First Amendment).

Extending First Amendment protection to video games certainly presents some thorny issues. After all, there are features of video games which are not terribly communicative, such as the manner in which the player controls the game. The plaintiffs in this case, however, complain about none of those non-expressive features. Instead, they argue that the video game, somehow, communicated to Carneal a disregard for human life and an endorsement of violence that persuaded him to commit three murders.

Because the plaintiffs seek to attach tort liability to the communicative aspect of the video games produced by the defendants, we have little difficulty in holding that the First Amendment protects video games in the sense uniquely relevant to this lawsuit. Our decision here today should not be interpreted as a broad holding on the protected status of video games, but as a recognition of the particular manner in which James seeks to regulate them through tort liability.

To say that the features of the defendants’ products of which the plaintiffs complain are protected by the First Amendment is not necessarily to say that attaching tort liability to those features raises significant constitutional problems. The plaintiffs’ argument is more nuanced: they do not seek to hold the defendants responsible merely for distributing their materials to anyone, but to young, impressionable children or, even more specifically, to Carneal. The protections of the First Amendment have always adapted to the audience intended for the speech.

Specifically, we have recognized certain speech, while fully protected when directed to adults, may be restricted when directed towards minors. See, e.g., Sable Communications v. FCC, 492 U.S. 115, 126, 109 S.Ct. 2829, 106 L.Ed.2d 93 (1989); Connection Distrib. Co. v. Reno, 154 F.3d 281, 291-92 (6th Cir.1998). We have also required, however, that such regulations be narrowly tailored to protecting minors from speech that may improperly influence them and not effect an “unnecessarily broad suppression of speech” appropriate for adults. Reno v.

American Civil Liberties Union, 521 U.S. 844, 874-75, 117 S.Ct. 2329, 138 L.Ed.2d 874 (1997); United States v. Playboy Entm’t Group, Inc., 529 U.S. 803, 812-13, 120 S.Ct. 1878, 146 L.Ed.2d 865 (2000); Connection Distrib., 154 F.3d at 292. Of course, the measure here intended to protect minors from the improper influence of otherwise protected speech is quite different from the regulations that we have countenanced in the past. Those regulations were the product of the reasoned deliberation of democratically elected legislative bodies, or at least regulatory agencies exercising authority delegated by such bodies.

It was legislative bodies that had demarcated what otherwise protected speech was inappropriate for children and that had outlined in advance the measures that speakers were required to take in order to protect children from the speech. In contrast, the question before us is whether to permit tort liability for protected speech that was not sufficiently prevented from reaching minors. At trial, the plaintiff would undoubtedly argue about the efficient measures that the defendants should have taken to protect the children.

But at the end of this process, it would be impossible for reviewing courts to evaluate whether the proposed protective measures would be narrowly tailored regulations. Who would.know what omission the jury relied upon to find the defendants negligent? Moreover, under the concept of negligence, there is no room for evaluating the value of the speech itself. The question before a jury evaluating James’s claim of negligence would ask whether the defendants took efficient precautions, in keeping their material from Carneal, that would be less expensive than the amount of the loss, the three deaths.

We cannot adequately exercise our responsibilities to evaluate regulations of protected speech, even those designed for the protection of children, that are imposed pursuant to a trial for tort liability.

Crucial to the safeguard of strict scrutiny is that we have a clear limitation, articulated in the legislative statute or an administrative regulation, to evaluate. “Whither our children,”as Appellant’s brief states at 3, is an important question, but their guidance through the regulation of protected speech should be directed in the first instance to . the legislative and executive branches of state and federal governments, not the courts. Our concerns about adequately evaluating the narrow tailoring of regulations wane if the underlying expression is completely unprotected.

Whether certain speech is not protected at all is a question that courts are regularly called upon to answer. James argues that to the extent that the content distributed by the defendants is at all expressive, it constitutes obscene material that is unprotected by the First Amendment. See Paris Adult Theatre I v. Slaton, 413 U.S. 49, 69, 93 S.Ct. 2628, 37 L.Ed.2d 446. (1973). It would be novel to argue that the video games and The Basketball Diaries are obscene.

As we understand James’s complaint, he complains that the video games and the movie are excessively violent — yet we have generally applied our obscenity jurisprudence, only to material of a sexual nature, appealing-only to prurient interest and lacking any other socially redeeming value. See Miller v. California, 413 U.S. 15, 23-24, 93 S.Ct. 2607, 37 L.Ed.2d 419 (1973) (‘We now confine the scope of [obscene .material] to works which depict or describe sexual conduct.”); United States v.

Thoma, 726 F.2d. 1191, 1200 (7th Cir.1984) (holding that because depictions of violence are not sexual, obscenity jurisprudence does not apply). With the sexually-oriented internet defendants, James has an arguable basis to allege that the material is unprotected speech under our conventional obscenity jurisprudence. Nevertheless, where the First Amendment concerns wane the most with the possibility of obscenity, the foreseeability problems loom the largest.

In his complaint, James contends that the sexual content displayed by the internet sites made “virtual sex pleasurable and attractive” for Carneal and caused him to undertake his violent actions. To us, however, James’s allegations regarding his exposure to violence seem much closer to the mark. After all, James does not seek to hold the internet defendants responsible for a sexual crime: Carneal never endeavored sexually to abuse Ms victims, just to kill them. With the movie and video game defendants, James contends that their material is excessively violent and constitutes obscene, non-protected speech.

We decline to extend our obscenity jurisprudence to violent, instead of sexually explicit, material. Even if we were to consider such an expansion, James’s arguments are not conceptually linked to our obscenity jurisprudence. The concept of obscenity was designed to permit the regulation of “offensive” material, that is, material that people find “disgusting” or “degrading.” See American Amusement Mach. Ass’n v. Kendrick, 244 F.3d 572, 574 (7th Cir.2001).

James’s argument, on the other hand, is that the violent content of these video games and the movie shapes behavior and causes its consumers to commit violent acts. This is a different claim than the obscenity doctrine, which is a limit on the extent to which the community’s sensibilities can be shocked by speech, not a protection against the behavior that the speech creates. Id. at 574-75 (holding that the law of obscenity does not apply to the argument that violence in video games causes players to commit violent acts).

This is not to say that protecting people from the violence that speech might incite is a completely impermissible purpose for regulating speech. However, we have generally handled that endeavor under a different category of our First Amendment jurisprudence, excluding from constitutional protection those communicated ideas and images that incite others to violence. Speech that falls within this category of incitement is not entitled to First Amendment protection. See R.A.V. v. City of St. Paul, 505 U.S. 377, 384-85, 112 S.Ct. 2538, 120 L.Ed.2d 305 (1992).

The Court firmly set out the test for whether speech constitutes unprotected incitement to violence in Brandenburg v. Ohio, 395 U.S. 444, 89 S.Ct. 1827, 23 L.Ed.2d 430 (1969). In protecting against the propensity of expression to cause violence, states may only regulate that speech wMch is “directed to inciting or producing imminent lawless action and is likely to incite or produce such action.” Id. at 447 (emphasis added). The violent material in the video games and The Basketball Diaries falls well short of this threshold.

First, while the defendants in this case may not have exercised exquisite care regarding the persuasive power of the violent material that they disseminated, they certainly did not “intend” to produce violent actions by the consumers, as is required by the Brandenburg test. Hess v. Indiana, 414 U.S. 105, 108, 94 S.Ct. 326, 38 L.Ed.2d 303 (requiring evidence that the speaker “intended” his words to produce violent conduct under the Brandenburg test).

Second, the threat of a person like Carneal reacting to the violent content of the defendants’ media was not “imminent.” Even the theory of causation in this case is that persistent exposure to the defendants’ media gradually undermined Carneal’s moral discomfort with violence to the point that he solved his social disputes with a gun. This glacial process of personality development is far from the temporal imminence that we have required to satisfy the Brandenburg test. Ashcroft v.

Free Speech Coalition, — U.S. -,-, 122 S.Ct. 1389, 1394, 152 L.Ed.2d 403 (2002) (“The mere tendency of speech to encourage unlawful acts is not a sufficient reason for banning it absent some showing of a direct connection between the speech and imminent illegal conduct.”); McCoy v. Stewart, 282 F.3d 626, 631 (9th Cir.2002) (holding that incitement is not imminent if it risks third-party violence at some “indefinite future time”).

Third, it is a long leap from the proposition that Carneal’s actions were foreseeable to the Brandenburg requirement that the violent content was “likely” to cause Carneal to behave this way. James contends that the Brandenburg test for speech inciting violence does not apply to depictions of violence, but instead only to political discourse advocating imminent violence. James suggests that the suppression of expression that is not advocacy, but does tend to inspire violence in its viewers or consumers, is governed by a less stringent standard.

Federal courts, however, have generally demanded that all expression, advocacy or not, meet the Brandenburg test before its regulation for its tendency to incite violence is permitted. See Dworkin v. Hustler Magazine, Inc., 867 F.2d 1188, 1199-1200 (9th Cir.1989) (applying the Brandenburg test to arguments that pornography causes viewers to engage in conduct that is violent and degrading to women); Herceg v.

Hustler Magazine, Inc., 814 F.2d 1017, 1024 (5th Cir.1987) (rejecting suggestion “that a less stringent standard than the Brandenburg test be applied in cases involving nonpolitical speech that has actually produced harm.”). Like the district court, we withhold resolution of these constitutional questions given the adequacy of the state law grounds for upholding the dismissal.

Attaching such tort liability to the ideas and images conveyed by the video games, the movie, and the internet sites, however, raises grave constitutional concerns that provide yet an additional policy reason not to impose a duty of care between the defendants and the victims in this case. B. Proximate Causation Even if this court were to find that the defendants owed a duty to protect James, Steger, and Hadley from Carneal’s violent actions, the plaintiffs likely have not alleged sufficient facts to establish the third element of a prima facie tort case: proximate causation.

The defendants argue that even if they were negligent, Car-neal’s intentional, violent actions constitute a superseding cause of the plaintiffs’ damages and sever the defendants liability for the deaths of Carneal’s victims. Generally, a third party’s criminal action that directly causes all of the damages will break the chain of causation. Yet, Kentucky courts have held that an intervening third-party criminal act will not be a superseding cause breaking the chain of causation if the act was reasonably foreseeable. See, e.g., Britton v. Wooten, 817 S.W.2d 448, 449-50 (Ky.1991).

In Britton, the plaintiff had been injured by a fire in his leased store. The fire started when trash piled next to the building by the landlord was ignited by an arsonist. The court held that the arsonist’s intervening act was not a superseding cause of the fire because the act was a reasonably foreseeable result of the landlord’s piling of flammable material next to the building. Ibid. The court reached the question of causation in Britton because of the landlord’s well established duty of care to protect its tenants from injury. See, e.g., Waldon v.

Housing Auth. of Paducah, 854 S.W.2d 777, 779 (Ky.Ct.App.1991) (holding landlord liable for failing to protect tenant from being shot by a third-party intruder). As we have noted above, the duty of the media defendants in this case is far from as firmly established. Outside of the landlord-tenant context, Kentucky courts are far more likely to determine that an intervening criminal act is a superseding cause. See Briscoe v. Amazing Prod., Inc., 23 S.W.3d 228, 229-30 (Ky.Ct.App.2000). Compare House v.

Kellerman, 519 S.W.2d 380 (Ky.1974) (confronting the question of superseding cause in the context of the well established duty of driver to protect his passenger). In Briscoe, the plaintiff was injured by a third party who battered her by spraying corrosive drain cleaner in her face. The plaintiff contended that the manufacturer of the drain cleaner was negligent and should have anticipated that it could be used as a weapon. The court held that batteror’s reaction to and use of drain cleaner as a weapon was sufficiently unforeseeable to constitute a superseding cause. Briscoe, 23 S.W.3d at 230.

A similar analysis is applicable here. Our determination regarding the idiosyncratic nature of Carneal’s reaction to the defendants’ media would likely compel us to hold that his action constitutes a superseding cause. We, however, need not reach this question because we have determined that the defendants did not owe a duty to protect the decedants. IV James also contends that the district court erred in dismissing his products liability claims.

In his complaint, James alleged that the defendants’ video games, movie, and internet transmissions constitute “products,” and their violent content “product defects.” Under Kentucky law, manufacturers, distributors, and retailers of “products” are strictly liable for damages caused by “defects” in those products. Clark v. Hauck Mfg. Co., 910 S.W.2d 247, 250 (Ky.1995); Restatement (Second) of Torts § 402A. If this theory of liability were to apply, James’s failure to establish a duty to exercise ordinary care to prevent the victims’ injuries (See

Part III.

A) would be irrelevant. Under strict liability, James would only be required to establish causation. James’s theory of product liability in this case is deeply flawed. First, and something none of the parties have mentioned, the “consumers or ultimate users” of the alleged products are not the ones claiming physical harm in this case. Restatement (Second) of Torts § 402A. Carneal was the person who “consumed” or “used” the video games, movie, and internet sites. Allegedly because of Carneal’s consumption of the products, he killed the victims in this case.

In early products liability law, courts had required privity between the final retailer of the product and the injured plaintiff for strict liability to attach. Eventually, courts broadened the class of plaintiffs who could avail themselves of strict liability to include those who consumed or ultimately used the product. Kentucky courts, but certainly not all courts, have extended the protection of products liability to “bystanders” who are injured by the product, but are not “using” or “consuming” the product. Embs v.

Pepsi-Cola Bottling Co. of Lexington, 528 S.W.2d 703 (Ky.1975) (holding that bystander could recover on strict liability basis for injuries caused by exploding soda bottle); Restatement (Second) of Torts § 402A, n.1 (noting the split in authority regarding whether bystanders may recover on a strict liability basis). Imposing strict liability for the injuries suffered in this case would be an extension of Kentucky’s bystander jurisprudence, as the decedents were not directly injured by the products themselves, but by Carneal’s reaction to the products.

We place this open question of Kentucky law aside as the parties apparently have. James has failed to demonstrate a prior requirement, that the video games, movies, and internet sites are “products” for purposes of strict liability. This was the basis on which the district court dismissed James’s products liability claims, holding that the video games, movie, and internet transmissions were not “products,” at least in the sense that James sought to attach liability to them. This court has already substantially resolved the question of Kentucky law presented. In Watters v.

TSR, 904 F.2d 378 (6th Cir.1990), this court held that “words and pictures” contained in a board game could not constitute “products” for purposes of maintaining a strict liability action. We cannot find any intervening Kentucky authority that persuades us that Watters no longer correctly states Kentucky law. James’s theory of liability, that the ideas conveyed by the video games, movie cassettes and internet transmissions, caused Carneal to kill his victims, attempts to attach product liability in a nearly identical way.

James argues that, at least in this case, we are not just dealing with “words and pictures.” Carneal, of course, had video game cartridges and movie cassettes. James argues that the test for determining whether an item is a product for purposes of Kentucky law is whether it is “tangible,” and that the cartridges and cassettes are “tangible.” As for the internet sites, James points us to a court that has held that “electricity” is a “product” for purposes of strict liability -under Kentucky law. See C.G. Bryant v. Tri-County Elec. Membership Corp., 844 F.Supp. 347, 352 (W.D.Ky.1994).

Internet sites are nothing more than communicative electrical pulses, James contends. James argues that there is no relevant difference between the internet transmissions and the electricity. And of course James is partially correct. Certainly if a video cassette exploded and injured its user, we would hold it a “product” and its producer strictly liable for the user’s physical damages.

In this case, however, James is arguing that the words and images purveyed on the tangible cassettes, cartridges, and perhaps even the electrical pulses through the internet, caused Carneal to snap and to effect the deaths of the victims. When dealing with ideas and images, courts have been willing to separate the sense in which the tangible containers of those ideas are products from their communicative element for purposes of strict liability. See, e.g., Winter v. G.P. Putnam’s Sons, 938 F.2d 1033, 1036 (9th Cir.1991); Jones v. J.B. Lippincott Co., 694 F.Supp. 1216, 1217-18 (D.Md.1988).

We find these decisions well reasoned. The video game cartridges, movie cassette, and internet transmissions are not sufficiently “tangible” to constitute products in the sense of their communicative content. y For all the foregoing reasons, we AFFIRM the district court’s dismissal of all James’s claims. . James directs us to several cases that, he argues, stand for the proposition that foreseeability of harm is a question of fact for the jury under Kentucky law, when reasonable minds can differ. Most of the cited cases, however, pertain to the question of foreseeability in the context of proximate causation.

Watts v. K, S, & H, 957 S.W.2d 233, 239 (Ky.1997) (“The question of foreseeable risk is covered by the usual instruction relating to proximate cause.”). Proximate cause, and its underlying foreseeability inquiry, is a question of fact for the jury. See

Part III.B. infra. The only other Kentucky case cited for this proposition has nothing to do with the common law of tort. See Finney Co. v. Monarch Constr. Co., 670 S.W.2d 857, 862-63 (Ky.1984) (discussing the concept of reasonably foreseeable reliance in a breach of contract case). James also directs us to Workman v. Columbia Nat. Res., 864 F.Supp. 638 (E.D.K.y.1994), which holds that foreseeability for determining the existence of a duty is a question of fact for the jury. The federal district court in Workman, however, erred in its

interpretation of Kentucky law on this issue, and we choose to follow the clear rulings of the Kentucky courts in Mullins and Sheehan. . Learned Hand's famous conception of negligence would hold conduct negligent if the harm that the conduct threatens to create, discounted by the ex ante probability of the harm's non-occurrence, outweighs the utility created by the conduct. See United States v. Carroll Towing Co., 159 F.2d 169 (2d Cir.1947). . Of course, James would continue to encounter proximate causation problems. See

Part III.B.