CONCURRING IN PART AND DISSENTING IN PART
BERNICE BOUIE DONALD,concurring in part and dissenting in part.
I concur in the majority’s decision to affirm the district court’s dismissal of Mitchell’s post-sale duty-to-warn and negligence claims. I dissent with respect to the majority’s finding that Mitchell’s pre-sale duty-to-warn claim fails as a matter of law.
I.
At the outset, the majority informs us that Robert Mitchell “resisted arrest” by “r[unning] from the car, t[aking] off on foot, b[reaking] into an abandoned house, and hid[ing] upstairs.” Maj. Op. at 225. Florid language aside, these facts are all technically true. The majority’s presentation of the facts, however, does not view the evidence in the light most favorable to Mitchell as the non-moving party — as we must at the summary-judgment stage. Shreve v. Franklin Cnty., Ohio, 743 F.3d 126, 132 (6th Cir.2014).
Taser does not dispute the following facts. On April 10, 2009, two Warren Police Department (“WPD”) officers initiated a traffic stop of a motor vehicle with an expired license plate. The stop occurred on a stretch of road between Warren, MI and Detroit, MI. Robert, who was 16 years old, was a passenger in that vehicle. Robert, “frightened” and “panicked,” jumped out of the passenger seat of the stopped car and ran from the scene. Robert’s reasons for fleeing remain unknown. Neither- he nor the other occupants of the vehicle had committed any crime, nor were they wanted for any crime. No evidence of any crime was ever recovered.
Robert ran into a nearby abandoned house and up a flight of stairs. Six police officers soon cornered Robert, ordering him to come downstairs. Robert immediately complied, and came downstairs with his hands extended and outstretched in front of him. He was “obviously unarmed.” Robert, an asthmatic, was “sweating profusely and breathing very heavily[,] ... out of breath and panting.” Robert — at-just 5'2" and 128 pounds — was substantially smaller than and clearly outnumbered by the six officers who surrounded him. He did not resist or attempt to flee again. At this point, the events in question were no longer tense, uncertain, or rapidly evolving. No use of force beyond the officers’ hand-holds was necessary to bring the incident to a peaceful close. Nonetheless, without any verbal *237warning or opportunity for Robert to respond, Officer Jesse Lapham instructed two other officers to step aside and deployed his Model X26 taser at the left side of Robert’s chest. The two electronically charged darts that were shot from the X26 penetrated the left side of Robert’s chest — five and one quarter (5 }4) inches apart — “creating a lethal trans cardiac vector.” Robert fell to the floor and officers handcuffed him. Thereafter, WPD officers noticed Robert was unresponsive, not breathing, and had no pulse. Resuscitation efforts failed and Robert was later pronounced dead. An autopsy listed Robert’s cause of death as “arrhythmogenic right ventricular cardiomyopathy (dyspla-sia) with use of a conducted energy device being a contributory factor.”
The differences between this presentation of the facts and the majority’s presentation of the facts may, at times, appear minor. But the language employed by the majority shifts the narrative — ie., our perception of Robert and the events leading to his death — in subtle but powerful ways. It is fair to question the majority’s inclination, however unintentional, to brand Robert as a miscreant — running from cars, taking officers on a foot chase, “breaking into” abandoned houses, and then hiding upstairs. What factual support does the majority have for the notion that Robert “tried to evade the officer’s grasp, and a struggle ensued,” thus prompting Officer Lapham to deploy his táser? Maj. Op. at 226. Nothing more than the deposition testimony of Officer Lapham himself. Again— with five other officers on the scene and deposition testimony from every single one of them — one questions the majority’s need to cite with approval the testimony of the one individual with the most to gain from painting the sixteen year old in the most negative light possible.
n.
The majority’s legal conclusions are also flawed. The majority correctly frames the issue presented: “Any liability for failure to warn in this case ... require[s] a showing that Taser ‘knew or should have known’ about the risk of cardiac arrest from an X26 chest shot based on the information available ‘at the time’ of sale — here August 17, 2006.” Maj. Op. at 226 (quoting Mich. Comp. Laws § 600.2948). But the majority, like the district court before it, exaggerates the question of whether Mitchell can prove actual knowledge and deemphasizes the equally material question of whether Mitchell can prove what Taser “should have known.”
A.
Taser contends that as of August 17, 2006, “no medical examiner, peer-reviewed literature[,] or plaintiff-retained expert report had EVER concluded that an X26 [conducted electrical weapon] directly induced VF or any fatal cardiac arrhythmia in any human.” The district court agreed, finding that the only evidence Mitchell proffered regarding whether Taser “knew of the risk of chest shots” were two studies published in the July 2006 volume of the Journal of the American College of Cardiology. Mitchell v. Taser Int’l, Inc., No. 09-11480, 2014 WL 3611632, at *1 (E.D.Mich. July 23, 2014). These two studies — conducted on sedated pigs — were referred to by the district court as the “Lakkired-dy/Tchou study” and the “Nanthakumar study.”1 Id. The district court summarized the studies as follows:
Both studies tested the effect of an X26 on cardiac rhythm in sedated pigs. In the Nanthakumar study, they analyzed 150 discharges, finding that dis*238charges to the chest area of the pigs often resulted in “myocardial [heart muscle] stimulation and capture [pacing].” After administering epineprhine [sic] to simulate stress, they induced ventricular fibrillation (“VF”) once in one pig.
The Nanthakumar study’s authors concluded that “[w]hen the discharge was vectored across the chest, electrical and mechanical capture [pacing] of the heart ensued.” “We also found that discharges away from the chest did not stimulate the heart or trigger arrhyth-mias.” Among other study limitations, the authors noted that “[t]he threshold for induction of VF in pigs may be lower than in humans, and the structural variation in the chest wall anatomy is another limitation with regard to extrapolating our model to humans.” ... The authors concluded that “[t]his study suggests that NIDs [neuromuscular incapacitating devices] may have cardiac risks that require further investigation in humans.”
In response to criticisms of the study, Nanthakumar stated: “The main point of our study was to demonstrate that electrical capture of the myocardium, under specific circumstances, can occur after neuromuscular incapacitating device (NID) discharge.... We agree that it is not possible to directly extrapolate our results to NID use in humans.... We did not state that NIDs cause ventricular fibrillation in humans, and we agree that we cannot conclude from our study that NID discharges cause ar-rhythmias in typical use. We hope that readers agree that our study does suggest the possibility that NIDs may, in some circumstances, cause cardiac capture, and that this possibility should at least be considered in future research in humans.”
The Lakkireddy/Tchou study considered “cocaine’s effects on Taser-induced ventricular fibrillation (VF) threshold in a pig model.” It concluded that “the presence of cocaine decreases the likelihood of NMI-induced VF.” ... They noted, however, that “[o]ur study is the first to describe capture of ventricular myocardium during application of NMI pulses.” This data “suggest the potential for induction of ventricular tachycardia [abnormal rapid heartbeat] in subjects with substrate for ventricular tachycardia....” and that “avoidance” of discharges near the heart “would greatly reduce any concern for induction of ventricular arrhythmias.” However, “[o]ur study showed that VF could not be induced using the standard 5 — s Ta-ser discharge applied to a pig’s body surface even at the most sensitive area tested.” ... “The results of our study and the few prior animal studies would suggest that NMI discharge at the standard 5 — s application is unlikely to cause life-threatening arrhythmias, at least in the normal heart.”
Id. at *1-2 (emphasis added) (citations omitted).
The district court concluded that neither study was sufficient to “establish TASER’s knowledge regarding the risk of VF or cardiac arrest in humans as a result of X26 shots to the chest.” Id. at *2. The court found that the Nanthakumar study merely “suggests that NIDs [neuromuscular incapacitating devices] may have cardiac risks that require further investigation in humans.” Id. (alteration in original) (internal quotation marks omitted). Likewise, the Lakkireddy/Tchou study found that a standard X26 discharge is “unlikely to cause life-threatening arrhythmias, at least in the normal heart.” Id. (internal quotation marks omitted). Further, Tchou testified that “because of our capture data, I would caution them [TASER] that there is *239some possibility that this could induce ventricular arrhythmias in people.” Id. (internal quotation marks omitted). The district court concluded that, at most, these studies only suggested “a theoretical risk as of August 2006, when TASER shipped the X26s to the Warren Police Department,” but this theoretical risk was “not sufficient to trigger a duty to warn under Michigan law.” Id. (internal citations omitted).
I believe the district court erred in framing its inquiry as whether Mitchell could establish that “TASER knew of the risk of chest shots.” Id. at *1. The question is not whether Mitchell had established actual knowledge. The question is whether there are any genuine issues of material fact regarding what Taser knew or should have known regarding the risk of VF from the X26. Mich. Comp. Laws § 600.2948(3). In other words, the query should have been what Taser knew, when Taser knew it, what Taser should have known, and when Taser shown have known it. In mischaracterizing Mitchell’s burden, the district court improperly held Mitchell to a higher burden of proof than that permitted by the summary-judgment standard. The district court then failed to draw all reasonable inferences in Mitchell’s favor as the non-moving party. Henderson v. Walled Lake Consol. Sch., 469 F.3d 479, 487 (6th Cir.2006).
B.
The majority would argue that, because our review of the district court’s summary judgment order is de novo, id. at 486, we may affirm the judgment of the district court on different grounds. This is, of course, correct. But the majority’s analysis, while distinct, fares no better than the district court’s analysis.
1.
The majority begins with a discourse on the historical and contemporary “field use of tasers.” This discussion draws primarily on reports, training, promotional, and testimonial materials from Taser International. It is hardly surprising that a company that manufactures and sells ta-sers would tout the virtues of the product it manufactures and sells. What is surprising — and improper at summary judgment — is that majority relies on these materials as one of “several reasons” no reasonable jury could find that Taser “knew or should have known of the risk of cardiac arrest from a X26 chest shot on August 17, 2006.” Maj. Op. at 226. Nowhere in her efforts to demonstrate a pre-sale duty to warn has Mitchell argued that the field use of tasers contributed to Taser’s actual or constructive knowledge of the risk of cardiac arrest from the X26 in August 2006, thus giving rise to pre-sale duty to warn. Nor has Mitchell argued that the complaint filed in William s v. Taser International, Inc. in January 2006 or the letter to the editor printed in the New England Journal of Medicine in September 2005 contributed to Taser’s actual or constructive knowledge. The majority’s discourse on this subject — an argument of the majority’s own creation, based largely on Taser’s propaganda materials and partly on independent research — serves no purpose other than to draw impermissible inferences in favor of Taser and stack the deck against Mitchell.
2.
Mitchell has argued that the two aforementioned studies gave Taser reason to know of the risk of cardiac arrest from an X26 chest shot. Both studies were published in the same August 2006 issue of the Journal of the American College of Cardiology. As the majority notes, the studies *240were “formally released a week before the City of Warren received its delivery of X26s but informally conveyed to Taser several months earlier.” Maj. Op. at 228. Also as the majority notes, Taser funded the Lakkireddy/Tchou study. Nonetheless, the majority concludes that the studies only found that although the X26 could “capture” a pig’s heartbeat, the studies “showed little risk that the standard charge could create cardiac arrest” — the cause of death in 16-year-old Robert Mitchell.
The majority neglects to mention that, in late 2005 or early 2006 — at least six months before the X26 sale to the WPD and more than three years before Robert’s death — Dr. Tchou and Dr. Lakkireddy met with Taser’s management, including CEO Patrick “Rick” Smith, to discuss the study’s progress. The two electrophysiol-ogists told TASER that it is “possible” for VF to occur when at least one taser dart lands “[i]n the region of the chest close to the heart.” Conversely, “it would be highly unlikely to have any directly induced arrhythmias from a TASER dart to ... parts of the body other than near the heart.” Dr. Tchou testified, moreover, that he informed Smith and other Taser representatives that “despite the fact that we did not induce ventricular fibrillation in any of the pigs, ... because of our capture data, I would caution them that there is some possibility that this could induce ventricular arrhythmias in people,” and it is “very clear from our data” that “the safety margin would increase if the dart locations were further away from the chest or the heart.”
Thus, although the majority is correct that the Lakkireddy/Tchou study “showed that [ventricular fibrillation] could not be induced using the standard 5-[second] [t]a-ser discharge applied to a pig’s body surface even at the most sensitive area tested,” Maj. Op. at 229, Taser was arguably aware — well in advance of the August 17, 2006, ship date — of the preliminary findings and recommendations of the study’s authors. These recommendations were later codified in the published study. But the majority, like the district court before it, selectively quotes the passages of the Lakkireddy/Tchou study that are favorable to Taser while ignoring the passages that are favorable to Mitchell. For instance, the Lakkireddy/Tchou study also noted, “Our data regarding myocardial capture, however, suggest the potential for induction of ventricular tachycardia,” and “[a]voidance of [darts near the heart] would greatly reduce any concern for induction of ventricular arrhythmias.” Likewise, the Nanthakumar study stated: “These findings suggest that there exists the possibility of serious ventricular arrhythmia during [X26] discharges in structurally normal hearts,” and “[o]ur findings of rapid ventricular stimulation with X26 discharge across the chest suggest a particular risk in individuals with pre-existing ... structural heart disease.”
The majority tells us, however, that • “[n]one of these studies established to a reasonable level of certainty that an X26 poses a risk of cardiac arrest in humans,” and that “more than ... a theoretical possibility” of cardiac arrest is required. Maj. Op. at 229. In reaching this conclusion, the majority improperly inserts itself into the role of the fact finder, an impermissible action at the summary-judgment stage. The Michigan statute “imposes a duty to warn that extends only to material risks not obvious to a reasonably prudent product user.” Greene v. A.P. Products, Ltd., 475 Mich. 502, 717 N.W.2d 855, 857 (2006). A “material risk” is one that presents “an important or significant exposure to the chance of injury or loss.” Id. at 860. The question, therefore, is whether the risk of cardiac arrest for an X26 chest shot is one *241that presents “an important or significant exposure to the chance of injury or loss”— not whether the risk is quantifiable or reaches some amorphous “reasonable level of certainty” satisfactory to the majority. Simply because a particular risk does not boast medical certainty does not automatically render the risk “speculative” or “theoretical.”2 Maj. Op. at 230. Drawing all reasonable inferences in favor of Mitchell, it is clear that there is medical consensus — as stated by Dr. Tchou, Dr. Nantha-kumar, and Dr. Zipes, Mitchell’s technical expert — that hitting the chest area with X26 darts creates a risk of cardiac arrest, while aiming away from the chest eliminates that risk altogether. Given the extreme consequences of cardiac arrest, and the relative ease with which that risk can be eliminated, a jury could find the cardiac capture risk documented by Dr. Tchou and Dr. Nanthakumar both “important” and “significant.” To conclude that the available evidence does not establish a genuine issue as to any material fact is disingenuous, as the majority in fact proceeds to resolve these issues in favor of Taser.
3.
The majority next charges that the arguments raised by Mitchell in support of reversal are “unpersuasive.” Maj. Op. at 231. For instance, the majority claims that Mitchell asserts that “we should not discount the pig studies because they are not human studies.” Id. at 231. Mitchell raises no such argument. Taser argues that “animal studies alone are insufficient to prove human causation under Michigan law.” For obvious ethical and safety reasons, Taser’s contention that a human study is obligatory before the company can reasonably be on notice of the risk of cardiac arrest from the X26 is simply absurd. More importantly, Taser’s demand for human testing is nothing more than an obvious attempt to undermine the findings of the Lakkireddy/Tchou and Nanthaku-mar studies — one of which Taser funded. Ultimately, Taser’s contention that the effect of an X26 on a pig does not exactly replicate the effect of an X26 on a human at most establishes that there are factual questions surrounding the studies. It does not establish, as a matter of law, that Taser had no knowledge of the potential risk to humans, and beliefs the conclusions of the study funded by and presented to Taser.
The majority next disregards Mitchell’s reliance on the Fourth Circuit’s decision in Fontenot v. Taser International, Inc., 736 F.3d 318 (4th Cir.2013). I agree with Mitchell’s contention that Fontenot illustrates how reasonable minds could disagree with the district court’s conclusions regarding the Lakkireddy/Tchou and Nan-thakumar studies.
The facts underlying Fontenot are similar to those in this case. In Fontenot, 17-year-old Darryl Turner died from cardiac arrest after a confrontation with police in which an X26 taser manufactured by Taser struck him in the chest. Id. at 321. The police officer who discharged the taser aimed the device at Turner’s chest based on training provided by the Charlotte Mecklenburg Police Department, which used instructional materials supplied by Taser. Id.
Tammy Lou Fontenot, Turner’s mother and the administrator of his estate, initi*242ated a products-liability action against Ta-ser in a North Carolina state court. Id. 321-22. In the complaint, Fontenot alleged that Taser negligently failed to warn users of the risk posed by the X26 taser and, in particular, to warn them to avoid-applying the taser’s electrical current near a subject’s heart. Id. at 322. Fontenot further alleged that Taser’s negligence was the proximate cause of Turner’s death. Id. A jury found in Fontenot’s favor, awarding her $10 million in compensatory damages, which amount the district court remitted to about $6.15 million before deducting certain offset amounts received by Fontenot, resulting in a final award of about $5.5 million. Id.
Taser appealed. The Fourth Circuit noted that the X26 “had been the subject of several academic studies,” and that Ta-ser “knew about these studies, in which researchers had concluded that the device posed a risk of ventricular fibrillation, a cause of cardiac arrest, especially when the electrical current from the taser was applied near the subject’s heart.” Id. In light of the studies’ findings, the Fourth Circuit found that Taser had “failed to warn taser users to avoid deploying the taser’s electrical current in proximity to the heart.” Id. The Fourth Circuit held:
Shortly after [Taser] issued the June 2005 Training Bulletin, [Taser] received the results of a [Taser]-funded study conducted by Dr.. Dhanunjaya Lakkired-dy concerning additional testing of the X26 taser. This study, which was published in the Journal of the American College of Cardiology, showed that the taser’s electrical pulses can “capture” cardiac rhythms, potentially leading to ventricular fibrillation. The study further noted that if users avoided striking the subject’s chest area with the taser’s darts, the risk of ventricular fibrillation would be reduced significantly.
[Taser] received the results of another study in 2006, which was conducted by Dr. Kumaraswamy Nanthakumar and was published in the same medical journal. Dr. Nanthakumar’s study likewise showed a risk of ventricular fibrillation in test animals when darts fired from the X26 taser lodged near the subject animal’s chest. Notably, the study showed that when the darts struck the animal in areas away from the chest, such as in the abdomen, the taser did not capture heart rhythms and, thus, using the taser in this manner avoided the risk of causing ventricular fibrillation.
These conclusions reached by Dr. Lakkireddy and Dr. Nanthakumar conflicted with [Taserj’s representations in its training materials that the X26 taser could not capture heart rhythms and was safe even when applied directly to a person’s chest. Nevertheless, as confirmed by [Taser]’s chief executive officer and the company’s vice president of training, [Taser] did not alter its training materials to warn users of the X26 taser that shots to a person’s chest could result in ventricular fibrillation, or that use of the taser near a person’s heart should be avoided based on that risk. Accordingly, up until the time of Turner’s death,- Captain Campagna and Officer Dawson continued to think that electrical current emitted by the X26 taser, even when applied near a person’s heart, did not affect heart rhythms or entail risks of cardiac arrest.
Id. 324-25.
The underlying events in Fontenot took place on March 20, 2008 — more than a year before Robert’s death. But the district court did not consider, or even discuss, the Fontenot decision — even though Mitchell’s cross-motion for partial summary judgment was premised on her as*243sertion that Fontenot had an issue-preclu-sive effect on this case. This was error. Even if the district court rejected the notion that issue preclusion applied, at the very least Fontenot is persuasive authority as to whether Taser may be held liable for harmful risks associated with chest shots from its X26 device under state-law theories of products liability and negligence.
The majority essentially finds that Fon-tenot is immaterial to the instant case because Fontenot was brought under North Carolina tort law and not Michigan tort law.3 This much is obvious. But Fontenot need not be legally on all fours with this case in order hold persuasive value. That is to say, the elements of a failure-to-warn claim under North Carolina law need not be precisely the same as the elements of a failure-to-warn claim under Michigan law in order for the North Carolina jury’s verdict to be relevant to whether genuine issues of material fact exist. In holding Taser liable on the facts of that case, the jury clearly would have had to make adverse factual determinations against Taser regarding what it knew and when it knew it. That adverse determination is directly relevant to the question of whether, in the present ease, Mitchell has established genuine issues of material fact regarding Ta-ser’s knowledge as of August 17, 2006. The fact that Fontenot proceeded all the way to trial — and that the Lakkired-dy/Tchou and Nanthakumar studies factored forcefully in proving Fontenot’s claims at that trial — strongly suggests that Mitchell has satisfied her burden at the summary-judgment stage.
Taser also argues that Fontenot is “ill-informed” because, in the Fourth Circuit case, Taser only appealed the damages award and did not contest the jury’s finding of liability. As an initial matter, it is difficult to comprehend how this argument does not undermine Taser’s cake here. If Taser only appealed as to damages, by definition Taser did not challenge the sufficiency of the evidence supporting the jury’s finding of liability against the company.
The majority, however, dismisses Fonte-not as a “2-1 decision” and “an outlier.” Maj. Op. at 2B2. The majority’s first point is ill-chosen. This case has also a 2-1 decision. I somehow doubt the majority finds its decision any less valuable because it resulted in a split decision. Finally, with respect to the majority’s contention that this case is an “outlier” as compared to decisions reached by the Fifth, Eighth, Ninth, and Eleventh Circuits, I would simply note that the majority makes no attempt to meaningfully discuss how these cases are legally and factually similar to the issue presented in this case. The majority’s failure to meaningfully discuss these cases is all the more glaring because the majority first seeks to distinguish Fon-tenot because it “involved North Carolina law,” and then cites with approval cases arising under Missouri, Mississippi, Arizona, California, and Georgia law. Id. Apparently, while the majority is permitted to find support for its position outside the state of Michigan, Mitchell cannot.
III.
For the foregoing reasons, I disagree with the majority’s finding that Mitchell’s *244pre-sale duty-to-warn claim fails as a matter of law. Genuine issues of material fact abound with respect to this claim. In deciding otherwise, the majority improperly usurps the role of the fact finder.
I respectfully dissent.
. For ease and consistency, this dissent will refer to these studies in the same way.
. The majority concedes as much in averring that "[cjourts do not demand scientific cer-‘ tainty” and "there are no certainties in science.” Maj. Op. at 230 (citations omitted). The majority's zero-sum analysis, however, suggests that because the risk of cardiac arrest from a X26 chest shot is not certain, that necessarily leads to the conclusion that the risk is speculative.
. The relevance of the majority’s observation that North Carolina law "does not limit the information the court could consider to studies published before the date of. sale," Maj. Op. at 231, is unclear. North Carolina law simply widens the window of information its courts may consider. Michigan law clearly creates a much smaller window. No one is suggesting, however, that this Court or that the district court should have considered information after the August 17, 2006, sale date that governs our inquiry. Accordingly, the majority’s observation is merely a distinction without a difference.