(This post contains spoilers for Avengers: Age of Ultron. You have been warned.)
In Avengers: Age of Ultron Tony Stark and Bruce Banner create Ultron, a brilliant but sociopathic artificial intelligence, which quickly obtains a physical form and wreaks havoc around the world. Although the Avengers eventually succeed in destroying Ultron, questions remain: are Stark and Banner in any way liable for the damage caused by Ultron’s actions? Could Ultron have been punished directly? And was Ultron’s own destruction at the hand of the Avengers legal? Today I’ll look at the first question. I’ll address the other two questions in future posts.
Under current US tort law, Stark and Banner may be liable for Ultron’s actions. First, they may be liable under an ordinary negligence analysis. Second, they may be liable under a strict liability analysis. Since Ultron’s creation and first harmful acts occurred in New York I will consider these questions from that perspective. To be sure, Ultron went on to cause many other injuries in many other jurisdictions, but for the sake of brevity I will limit myself to what happened at Stark Tower.
In New York, “a plaintiff must demonstrate the existence of a duty of care owed to the plaintiff, a breach of that duty, and that the breach of such duty was a proximate cause of his or her injuries.” Miglino v. Bally Total Fitness, 92 A.D.3d 148, 159 (N.Y. App. Div. 2011). In this case, Stark invited friends to his home (which I assume he owns), where they were injured. He owed those present on his property a duty of reasonable care. Tagle v. Jakob, 97 N.Y.2d 165, 166 (2001) (noting the abolition of the distinction between trespassers, licensees, and invitees). He arguably breached that duty of care when he experimented with an ill-understood, incredibly powerful alien artifact.
(For this question I will focus on Stark rather than Stark and Banner because Stark’s duty of care is more easily established. Banner could be liable as a joint tortfeasor of some kind, but I don’t want this post to get too complicated.)
Two questions remain in the analysis: first, did Stark take reasonable care to avoid injury? And second, was his lack of reasonable care (if any) the proximate cause of the injuries?
“The objective, reasonable person standard in basic traditional negligence theory, however, necessarily takes into account the circumstances with which the actor was actually confronted when the accident occurred, including the reasonably perceivable risk and gravity of harm to others and any special relationship of dependency between the victim and the actor.” Bethel v. NYC Transit Authority, 92 N.Y.2d 348, 353 (1998). In this case we can rule out special relationships and instead focus on the actual circumstances and reasonably perceived risk and gravity of harm.
The actual circumstances were that Stark conducted this experiment in his home with guests present, who were unaware of the experiment and who would, as Stark knew, disapprove of it and thus, at a minimum, have not been present had they known of its existence. Further, Stark and Banner knew that there were significant risks in experimenting with the scepter, and a reasonable person would have perceived both those risks and the gravity of the potential harm because of the known injury that the scepter had caused in the past.
Given that the risk was high and the gravity of harm was high, did Stark take reasonable care to avoid injury? I don’t think he did. He concealed the experiment from his guests, instead drinking with them while the experiment proceeded, and that after 3 days of apparently uninterrupted work. The only consolation is that he left the experiment attended by Jarvis, who is ordinarily a very capable assistant, but nonetheless as we saw he was unable to contain Ultron. I believe someone taking ordinary care would not have left the experiment while it was in progress, would not have invited uninformed guests to the premises, and would not have consumed alcohol under those circumstances.
Finally, then, was Stark’s lack of care the foreseeable, proximate (i.e. legally responsible) cause of the injuries? Or were Ultron’s own actions a superseding cause that broke the chain of responsibility?
This is a difficult question. Would a reasonable person foresee that experimenting with an alien artificial intelligence could result in malevolent robots attacking the guests at the party? It’s possible, but this is quite possibly the weakest part of the argument for negligence liability. Stark and Banner did not intentionally place the artificial intelligence in a robot body. Ultron’s disembodied program caused the creation of a body and directed other Iron Legion robots to attack the guests. The distance in causal steps makes it harder to argue that a reasonable person would have foreseen that risk.
B. Superseding Causes
“Where the acts of a third person intervene between the defendant’s conduct and the plaintiff’s injury, the causal connection is not automatically severed. In such a case, liability turns upon whether the intervening act is a normal or foreseeable consequence of the situation created by the defendant’s negligence.” Derdiarian v. Felix Contracting Corp., 51 N.Y.2d 308, 314 (1980).
This in turn raises the question of whether Ultron was a third person for legal purposes. As discussed in The Law of Superheroes, the law does not currently recognize machines as legal persons, and the existing legal framework suggests that only Congress can extend legal personhood to a non-human. Absent such a law, Ultron would likely not be considered a person.
But if Ultron were considered a person, would his criminal, tortious actions sever liability for Stark? “Where third-party criminal acts intervene between defendant’s negligence and plaintiff’s injuries, the causal connection may be severed, precluding liability.” Bell v. Board of Educ. City of N.Y., 90 N.Y.2d 944, 945 (1997). However, if the criminal acts are a reasonably foreseeable result of circumstances created by the defendant, then liability is not severed. Id. In this case, if it was foreseeable that the experiment could create a malevolent robot, then that robot’s criminal actions are likewise a foreseeable result of circumstances created by the defendant, since the defendant’s actions literally created the robot and provided it with an immediate opportunity to cause injury.
Overall, the case for negligence is a little weak. Foreseeability is a difficult argument to make because the circumstances are so extreme. It’s helped somewhat by past experience with the scepter and the known lethality of Stark’s robots, but it’s still a bit of a stretch. What about strict liability?
II. Strict Liability
Strict liability does away with questions of duty and reasonable care. Instead, when strict liability applies, all that matters is causation and injury. In New York “those who engage in activity of sufficiently high risk of harm to others, especially where there are reasonable even if more costly alternatives, should bear the cost of harm caused the innocent.” Doundoulakis v Town of Hempstead, 42 N.Y.2d 440, 448 (1977). Thus, if experimenting with an alien artificial intelligence in a residence in Manhattan carries a “sufficiently high risk of harm to others” (i.e. is an “abnormally dangerous activity”), then Stark is liable for any resulting damage, even if he owed no duty of care and took extraordinary care.
There is no hard and fast rule for determining whether an activity is abnormally dangerous, but there are some commonly used factors: “(a) existence of a high degree of risk of some harm to the person, land or chattels of others; (b) likelihood that the harm that results from it will be great; (c) inability to eliminate the risk by the exercise of reasonable care; (d) extent to which the activity is not a matter of common usage; (e) inappropriateness of the activity to the place where it is carried on; and (f) extent to which its value to the community is outweighed by its dangerous attributes”. Id.
(a) Because of past experience with the scepter, this seems to be satisfied. Banner in particular initially argued against the idea of experimenting with the AI. And presumably Thor and the other Avengers would have been opposed.
(b) Again, past experience with the scepter, the skepticism of others, and the grievous harm that actually resulted suggests that the harm was likely to be great.
(c) The insidious, malicious nature of the AI suggests that reasonable care could not have eliminated the risk. If two of the most brilliant scientists in the world, with essentially unlimited resources at their disposal, could not eliminate the risk, then it probably could not be eliminated. Indeed, eliminating the risk (in the form of creating the Vision) ultimately required the employment of the power of a god (well, a godlike alien, but you get the idea).
(d) This kind of goes without saying.
(e) Again, this all happened in a residence while guests were present, some of whom did not have superpowers. Seems like an inappropriate place to me.
(f) This is a bit of a sticking point. Banner (and the other Avengers) would have argued that it wasn’t worth the risk. Stark had an argument otherwise. It’s a debatable point. But at most I’d say it’s weakly arguable that it’s worth the risk.
All told I think there’s a good case for strict liability here. Messing around with phenomenally powerful, ill-understood alien artifacts in a residential area for dubious reasons seems like a good example of an abnormally dangerous or ultrahazardous activity. So probably don’t do that, unless you’ve got Stark money.