1 Deepmind Health, https://deepmind.com/applied/deepmind-health/ (last visited October 17, 2017) [https://perma.cc/VZ5D-QJAY].

2 Inner – Eye Medical Imaging AI to Empower Clinicians, https://www.microsoft.com/enus/research/project/medical-image-analysis/ (last visited October 17, 2017) [https://perma.cc/PG9F-7SED].

3 IBM Watson, https://www.ibm.com/watson/ (last visited October 17, 2017) [https://perma.cc/LVM5-4MNY].

4 John Markoff, Computer Wins on ‘Jeopardy!’: Trivial, It's Not, N.Y. TIMES, Feb. 16, 2011, http://www.nytimes.com/2011/02/17/science/17jeopardy-watson.html?pagewanted=all.

5 See id.

6 Alex Hogan & Ike Swetlitz, How Does Watson for Oncology Work?, STAT (Sept. 5, 2017), http://content.jwplatform.com/previews/7jAqNnX0-jEuQjxp9 [https://perma.cc/3GGZ-DKUU].

7 IBM Watson Health, https://www.ibm.com/watson/health/oncology-and-genomics/oncology/ (last visited October 18, 2017) [https://perma.cc/M7CX-D3EE].

8 Hogan & Swetlitz, supra, note 6.

9 Id.

10 Id.

11 Id.

12 Id.

13 Id.

14 Id.

15 Id.

16 Id.

17 Id.

18 See Ariel Bleicher, Demystifying the Black Box that is AI, Sci. Am., Aug. 9, 2017, https://www.scientificamerican.com/article/demystifying-the-black-box-that-is-ai/ [https://perma.cc/LN7J-7S2J].

19 Memorial Sloan Kettering Cancer Center – Watson Oncology, https://www.mskcc.org/videos/mskcc-andibm-collaborate-applying-watson-technology-help-oncologists (last visited October 18, 2017) [https://perma.cc/9CSS-YK5Y].

20 Hogan & Swetlitz, supra note 6.

21 Casey Ross & Ike Swetlitz, IBM Pitched its Watson Supercomputer as a Revolution in Cancer Care. Its Nowhere Close, STAT (September 5, 2017), https://www.statnews.com/2017/09/05/watson-ibm-cancer/?utm_source=STAT+Newsletters&utm_campaign=9406bd149a-EMAIL_CAMPAIGN_2017_09_06&utm_medium=email&utm_term=0_8cab1d7961-9406bd149a-150109745 [https://perma.cc/F3YA-CF59].

22 Hogan & Swetlitz, supra, note 6.

23 Id.

24 See About Clinical Trials, https://www.nhlbi.nih.gov/health-topics/about-clinical-trials (last visited March 5, 2018) [https://perma.cc/G763-XBSN] (explaining that “clinical trials are research studies that explore whether a medical strategy, treatment, or device is safe and effective for humans. These studies also may show which medical approaches work best for certain illnesses or groups of people. They also can help health care decisionmakers direct resources to the strategies and treatments that work best.”)

25 Ross & Swetlitz, supra, note 21.

26 Id.

27 This scenario is true in America, but oncologists who work in under-resourced countries may rely exclusively on the recommendations WO provides. Ross & Swetlitz, supra, note 21.

28 A concordance rate is the proportion of observations on which two observers agree (i.e. two doctors see 100 patients and both doctors observe that the same 10 patients have a gallop and the same 75 others do not). The concordance rate is calculated by adding the number of times the doctors agreed and dividing that by the total number of subjects. The math would be (10+75)/100, for a concordance rate of 85%. Stephen B. Hulley, et al., Designing Clinical Research 200 (Lippincott Williams & Wilkins ed., 3rd ed. 2007). In the context of WO, its concordance rate (percentage of agreeance) would compare the number of times WO and an oncologist agree on the course of treatment to pursue in relation to a given cancer patient.

29 Ross & Swetlitz, supra, note 21.

30 See id.

31 Id.

32 Id.

33 Id.

34 S.P. Somashekhar et al., Abstract, Am. Soc'y of Clinical Oncology, (2017).

35 Choi, Youn I. et al., Concordance Rate between Oncologists and Watson for Oncology among Patients with Advanced Gastric Cancer: Early, Real-World Evidence in Korea, 19 Can. J. of Gastroenterology & Hepatology, 1, 2–3 (2019)Google Scholar (noting that discordance can occur due to complex medical histories and oncologist preferences for chemotherapy as a course of treatment).

36 A patient injured by an artificially intelligent medical device could proceed under a theory of vicarious liability, medical malpractice, or products liability. The argument advocating to use products liability will be addressed below.

37 Silla Brush, et al., How a Mystery Trader with an Algorithm May Have Caused the Flash Crash, Bloomberg News (April 22, 2015), https://www.bloomberg.com/news/articles/2015-04-22/mystery-trader-armed-with-algorithms-rewrites-flash-crash-story (https://perma.cc/RC6Z-52KB).

38 Id.

39 Stoll, Hans R., Electronic Trading in Stock Markets, 20 J. of Econ. Persp. 154, 171 (2006)Google Scholar (discussing the phenomenon known as spoofing).

40 See generally Kirilenko, Andrei A., et al., The Flash Crash: High-Frequency Trading in an Electronic Market, 72 J. of Fin. 967–998 (2017)CrossRefGoogle Scholar (explaining that high frequency traders (HFTs) were only one of many causes).

41 Graham Bowley, Stock Swing Still Baffles, With an Ominous Tone, N.Y. Times (August 22, 2010), http://www.nytimes.com/2010/08/23/business/23flash.html [https://perma.cc/XQ9V-L8H7].

42 Press Release, Department of Justice, Futures Trader Charged with Illegally Manipulating Stock Market, Contributing to the May 2010 Market ‘Flash Crash’ (April 21, 2015).

43 Allain, Jessica S., From Jeopardy to Jaundice: The Medical Liability Implications of Dr. Watson and Other Artificial Intelligence Systems, 73 LA. L. Rev. at 1049, 1051 (2013)Google Scholar.

44 Id.

45 Id. at 1072-77.

46 Id at 1068; Under the products liability regime, manufacturers have a duty to warn physicians of the medical device's potential dangers. Hamilton-Piercy, Meghan, Cybersurgery: Why the United States Should Embrace this Emerging Technology, 7 J. High Tech. L. 203, 212 (2007)Google Scholar. By warning the treating physician, the manufacturer is effectively discharging any duty it would have owed to the patient. Banker v. Hoehn, 718 N.Y. S.2d 438, 440 (N.Y. App. Div. 2000)Google Scholar (citations omitted); Ehlis v. Shire Richwood, Inc., 367 F.3d 1013, 1019 (8th Cir. 2004)Google Scholar;

47 Software is generally not within the scope of products liability because courts consistently hold that software is not a product. Childers, Seldon J., Don't Stop the Music: No Strict Products Liability for Embedded Software, 19 U. Fla. J. of L. & Pub. Pol'y, 128 (2008)Google Scholar. Id. at 1069-1070.

48 Allain, supra, note 43 at 1075.

49 Id.

50 Id. at 1075-76 (explaining that “the enterprise would include Watson, his owner, and all the physicians that were part of the treatment plan that led to injury” and that “WO should be protected like physicians under state medical malpractice regimes, including limits on liability.” Thereafter, the court would proceed by assessing causation and then damages.)

51 Id. at 1051.

52 Id. at 1074, 1076; Because WO is one of several members of the physician team, WO would never be solely responsible for a plaintiff's injuries. Thus, each treating physician would be responsible for damages equivalent to their liability.

53 Id. at 1073-1074.

54 Vladeck, David C., Machines without Principals: Liability Rules and Artificial Intelligence, 89 Wash. L. Rev. at 117, 120-21 (2014)Google Scholar.

55 Id. at 128.

56 Id.

57 Id. at 146.

58 Id. at 146-147.

59 See generally Sjur Dyrkolbotn, A Typology of Liability Rules for Robot Harms, in A World with Robots (Maria Isabel Aldinhas Ferreira et al. eds., 2017).

60 Id.

61 Id at 130.

62 Id.

63 Id.

64 Vladeck, supra, note 54 at 142 (referring to truly autonomous artificially intelligent systems); Allain, supra, note 43 at 160-161 (referring to artificially intelligent systems that independently practice medicine).

65 Dyrkolbotn, supra, note 59 at 120-121 (noting that when complex intelligence systems are involved, it is difficult for plaintiffs to produce evidence that supports their damage claims).

66 Different states require plaintiffs to prove different things when alleging a product's design is flawed. For instance, Texas requires a plaintiff prove by a preponderance of the evidence that there is a safer alternative design that reasonably would have prevented or significantly reduced the risk of the claimant's personal injuries without impairing the product's utility and was economically and technically feasible. See Allstate Lloyds Co. v. Marvin Lumber and Cedar Corp., 2006 WL 2506776 at *6 (Tex. App. 2006).

67 Ross & Swetlitz, supra, note 21.

68 With respect to WO, Denmark reported a concordance rate of 33 percent while the American Society of Clinical Oncology reported a 96 percent concordance rate. Ross & Swetlitz, supra, note 21.

69 See Kenneth B. Makmberg, The Role of Government Regulation and Leadership in Increasing Sustainability, Am. Soc'y for Pub. Admin. (2013), https://patimes.org/role-government-regulation-leadership-increasing-sustainability/ [https://perma.cc/8M8W-NM5W].

70 Federal Food, Drug, and Cosmetic Act, Pub. L. No. 115-90 (codified in scattered sections of 21 U.S.C. Ch. 9).

71 IBM, supra at 1, (noting that WO is a cognitive computing decision support system).

72 Tsang, Lincoln et al., The Impact of Artificial Intelligence on Medical Innovation in the European Union and the United States, 29 Intell. Prop. & Tech. L. J. 3, 10 (2017)Google Scholar.

73 Id. at 7.

74 Id.

75 Tsang, supra note 72, at 8 (noting that IBM's Watson is a CDSS device).

76 21st Century Cures Act, Pub. L. No. 114-255 § 3060 (codified at 21 USC § 360(j)).

77 Jackson, Vicki C., Suing the Federal Government: Sovereignty, Immunity, and Judicial Independence, 35 Geo. Wash. Int'l L. Rev. 521, 525 (2003)Google Scholar.

78 Id.

79 Restatement (Third) of Torts § 13 (Am. Law Inst. 2000).

80 See Id.; 65 C.J.S. Negligence § 150 (2011).

81 Southwick, Arthur F., Vicarious Liability of Hospitals, 44 Marq. L. Rev. 153, 153 (1960)Google Scholar.

82 Id.

83 See Allain, supra, note 43 at 1066-1067 (a discussion of how a vicarious liability suit would proceed if WO were considered an employee of the institution in question.)

84 Id.

85 Southwick, supra note 81 at 154.

86 Only four states have waived their immunity from liability. Ala. Comp. Law Ann. §56-7-1, 56-7-7 (Supp. 1958). Ky. Rev. Stat. Ann. §44.070 (Supp. 1958) (Board of Claims hears claims against the state for negligent acts and may pay claimant up to $10,000). N.Y. Court of Claims Act Section 8. N.C. Gen. Stat. §§143-291 (1958) (North Carolina Industrial Commission operates as a court to hear tort claims and may award damages up to $10,000).

87 Although the federal government itself is immune from suit if a hospital it owns is sued, the doctors who practice at the hospital in question may be sued under the Federal Torts Claims Act. Southwick, supra note 81, at 154.

88 Id.

89 Steven E. Pegalis, American Law of Medical Malpractice § 6:19-6:20 (2017), available at Westlaw.

90 Allain, supra note 43, at 1065.

91 The plaintiff must prove the oncologist committed medical malpractice because vicarious liability is predicated upon the oncologist's wrongdoing. See Southwick, supra note 81, at 153.

92 The black box effect of artificial intelligence machines refers to the inability of humans to perceive or understand the reasoning and decision making of AI. See supra note 18.

93 See Mracek v. Bryn Mawr Hosp., 610 F.Supp. 2d 401, 403 (declining to find in favor of the plaintiff for products liability, strict malfunction liability, and negligence despite the fact that the surgical AI device displayed error signs).

94 Id. at 405 (finding that the fact that the surgical device repeatedly shut down and flashed error signs was not obvious proof of a defective design, and thus such a conclusion needed to be made by an expert and not a jury.)

95 Allain, supra note 43, at 1061.

96 See Id.; For example, physicians have a duty to employ their best judgment, Pegalis, supra note 89, at § 3:6, to take proper medical history, Id. at § 3:9, to conduct careful and proper physical examinations, Id. at § 3:10, to properly use laboratory and ancillary procedures, Id. at § 3:11, to make a proper diagnosis, Id. at § 3:12, to refer to a specialist, Id. at § 3:13, and to not abandon a patient, Id. at § 3:15. Some states also define a physician's duty by what is customary, instead of asking what a reasonably prudent physician would have done. See Moffett, Peter MD and Moore, Gregory MD, JD, The Standard of Care: Legal History and Definitions: the Bad and Good News, 12 Western J. Emergency Med. 109, 109-10 (2012)Google Scholar.

97 See e.g., Kuznar v. Raksha Corp., 750 N.W.2d 121, 128 (Mich. 2008) (holding that a pharmacist could not be held liable under medical malpractice because there was no physician-patient relationship); Oblachinski v. Reynolds, 706 S.E.2d 844, 846 (S.C. 2011) (holding that only a patient can bring a medical malpractice claim against a physician, as opposed to third parties affected indirectly); and Estate of French v. Stratford House, 333 S.W.3d 546, 555 (Tenn. 2011) (holding that a physician-patient relationship is a prerequisite to a claim under the Tennessee Medical Malpractice Act).

98 Pegalis, supra note 89 at §§ 4:1 and 4:10.

99 The standard by which the physician should be judged is the degree of knowledge and concomitant medical or surgical skill that a physician under the same or similar circumstances should “reasonably” possess. Pegalis, supra note 89 at §§ 4:1.

100 See Allain, supra, note 43, at 1063. If WO is not treated as a consulting physician, then the suit would proceed as a regular medical malpractice claim as described in the paragraph preceding this footnote.

101 Granade, Phyllis Forrester, Medical Malpractice Issues Related to the Use of Telemedicine – An Analysis of the Ways in Which Telecommunications Affects the Principles of Medical Malpractice, 73 N.D. L. Rev. 65, 68 (1997)Google Scholar.

102 See Irvin v. Smith, 31 P.3d 934, 941 (Kan. 2001) (determining that a consulting physician who only gives an informal opinion to a treating physician owes no duty to the patient); St. John v. Pope, 901 S.W.2d 420, 424 (Tex. 1995) (holding that a consulting physician who had recommended that the patient be transferred to a different facility for treatment and had not agreed to treat the patient did not owe the patient a duty); and Hill v. Kokosky, 463 N.W. 2d 265, 267 (Mich. Ct. App. 1990) (holding that the consulting physicians did not establish a physician-patient relationship with the plaintiff and, therefore, owed no duty to her).

103 Allain, supra note 43 at 1064.

104 See IBM, The Science Behind an Answer, YouTube (Jul. 18, 2011) http://www.youtube.com/watch?v=DywO4zksfXw [https://perma.cc/MS6N-DBL8].

105 See Pegalis, supra note 89.

106 Suing the FDA is a legal impossibility as the federal government is immune from liability. See generally Jackson, supra note 77.

107 See supra note 91 and the accompanying text.

108 Allain, supra note 43 at 1064 and 1069.

109 See generally Bleicher, supra, note 18 (“Therein lies today's AI conundrum: The most capable technologies—namely, deep neural networks—are notoriously opaque, offering few clues as to how they arrive at their conclusions. “[Deep neural nets] can be really good but they can also fail in mysterious ways,” says Anders Sandberg, a senior research fellow at the University of Oxford's Future of Humanity Institute.”).

110 Software is not within the scope of products liability. Childers, supra note 47 at 128.

111 The doctrine of res ispa loquitur is a theory that allows courts to presume that the plaintiff's harm is a result of the negligence of the defendant when (1) the event is of the type that does not normally occur in the absence of negligence, (2) other potential causes are sufficiently eliminated, and (3) the negligence in question is within the scope of the defendant's duty to the plaintiff. See Restatement (Third) of Torts § 17 (2017). Some states require the harm-causing instrumentality be in the exclusive control of the defendant or the defendant's agents. 65A C.J.S. Negligence § 871 (2017).

112 See supra note 21.

113 “The expert-witness arms race in which we find ourselves makes litigation more expensive, more complicated, and more protracted than it would otherwise be.” Shaughnessy, Robert J., Dirty Little Secrets of Expert Testimony, 33 Litig. 47, 52 (2007)Google Scholar.

114 Restatement (Third) of Torts § 1 (2017).

115 Id. at §2.

116 The learned intermediary doctrine was originally developed to apply to pharmaceuticals and was expanded to include medical devices. McLean, Thomas R., Cybersurgery–An Argument for Enterprise Liability, 23 J. Legal Med. 167, 183-184 (2002)CrossRefGoogle Scholar.

117 Hamilton-Piercy, Meghan, Cybersurgery: Why the United States Should Embrace This Emerging Technology, 7 J. High Tech. L. 203, 212 (2007)Google Scholar.

118 A product is defective in design when the foreseeable risks of harm posed by the product could have been reduced or avoided by the adoption of a reasonable alternative design by the seller or other distributor, or a predecessor in the commercial chain of distribution, and the omission of the alternative design renders the product not reasonably safe. Restatement (Third) of Torts § 2 (1998).

119 The Restatement Third of Torts has been interpreted as eliminating strict products liability. Godoy ex. Rel. Gramling v. E.I. du Pont de Nemours and Co., 768 N.W.2d 674, 689-690 (Wis. 2009).

120 63 Am. Jur. 2D Products Liability § 8 (2017).

121 Talley v. Danek, Inc., 179 F.3d 154, 162 (4th Cir. 1999) (declining to find the Dyna-Lok device was defectively designed).

122 Id. at § 6.

123 Restatement (Third) of Torts § 2 (2017).

124 Am. Jur. supra note 120 at § 15.

125 Id.

126 See e.g. Fairley v. American Hoist & Derrick Co., 640 F.2d 679, 681 (5th Cir. 1981) (holding the plaintiff must carry the burden of establishing by a preponderance of the evidence that the product when sold was in a defective condition, unreasonably dangerous to the user or the consumer; the product must reach the user or consumer without substantial change in the condition existing at the time of sale; and the defective condition must be a substantial element in the cause of the injury); Stanton by Brooks v. Astra Pharm. Prod., Inc., 718 F.2d 553, 571 (3d Cir. 1983) (proving a pharmaceutical drug was defective because it was not compliant with FDA regulations at the time it left the manufacturer to be sold).

127 Am. Jur. supra note 120 at § 14.

128 Alternative causes of a product's defect could include improper handling, Id. at § 16, the lapse of time between the manufacturer of a product and its sale, Id. at Products Liability § 17, or the long-continued use of a product. Id. at Products Liability § 18.

129 See supra note 118 (“Moreover, the nature of the product itself may be a factor supporting an inference that a defect which caused an accident or injury existed in the hands of the defendant.”)

130 See supra, note 7 (“Watson for Oncology helps physicians quickly identify key information in a patient's medical record, surface relevant evidence and explore treatment options.”)

131 Darryl v. Ford Motor Co., 440 S.W.2d 630, 632 (Tex. 1969).

132 Id.

133 See supra, note 118.

134 Restatement (Third) of Torts § 3 (2017).

135 GTE Corp. Allendale Mutual Insurance Co., 372 F.3d 598, 610 (2004)Google Scholar (finding that a manufacturer's design is not reasonable simply because it complies with industry best practices and regulatory requirements due to the fact that the manufacturer ignored a shortcoming in the design that remediation efforts revealed.)

136 Restatement (Third) of Torts § 2, cmt. b (2017). For example, a plaintiff could use evidence of another's injury to bolster the fact that she was also injured by the same product. Peters v. General Motor Corp., 200 S.W.3d 1, 11 (Mo. Ct. App. 2006) (finding that evidence of seven other accidents that were not caused by driver acceleration could sufficiently prove that the plaintiff's accident was caused by the same defect; however, evidence of the seven other accidents was not admissible to demonstrate defective design because the seven other plaintiffs did not eliminate other alternative causes).

137 See id. Proving the design of a product is defective with circumstantial evidence, as opposed to offering evidence of a feasible alternative design, most frequently applies to manufacturing defects but is applicable to design defect cases. See See id. at cmt. b.

138 See id.

139 Hulley, supra note 28.

140 Id. It is also important to note that, because WO makes three tiers of treatment recommendations, a concordance rate is calculated for each of the red, yellow, and green recommendations. See e.g. Choi et al., supra note 35, at 2.

141 Ross & Swetlitz, supra note 21.

142 Krippendorff, Klaus, Bivariate Agreement Coefficients for Reliability of Data, in Soc. Methodology 139 (Wiley, vol. 2, 1970)Google Scholar.

143 WO's concordance rate ranges between 33% and 96%, and it is true that 96% is objectively high, but the lack of research into WO's concordance rates makes it unlikely that the manufacturer could prove the high concordance rate means WO is reliable; not at least without revealing information that could help the plaintiff prove her case.

144 Restatement (Third) of Torts § 3, cmt. a (2017)

145 See id. at § 3 (2017).

146 Susan G. Komen Breast Cancer Foundation, Factors that Affect Prognosis and Treatment, https://ww5.komen.org/BreastCancer/FactorsthatAffectTreatmentOptions.html [https://perma.cc/YXX7-8RGS].

147 Products liability cases are oftentimes won and fought through battling expert testimony. See e.g., In re Lipitor (Atorvastatin Calcium) Marketing, Sales Practices, and Products Liability, 227 F.Supp.3d 452, 456 (D. S.C. 2017) (discussing whether an expert's testimony should be admitted.); Vladeck, supra note 52, at 138-139 (discussing the hurdles an expert would need to overcome to prove a product is defective under the Restatement (Third) Torts.)

148 Oncologists are not bound to accept the recommendation WO provides, but it likely is improbable the oncologist would pursue another course of treatment other than the one recommended. WO is valuable because it saves oncologists from having to perform hours or even days of research, so if the course of treatment is not obviously flawed then it seems the oncologist will pursue the recommendation offered.

149 Restatement (Third) of Torts § 2 (2017) (“A product is defective in design when the foreseeable risks of harm posed by the product could have been reduced or avoided.”)

150 Id. at cmt. b.

151 Horst v. Deere & Co., 769 N.W.2d 536, 560 (Wis. 2009).

152 Fed. R. Evid. 608.

153 Cory Doctorow, Watson for Oncology isn't an AI that Fights Cancer, it's an Unproven Mechanical Turk that Represents the Guesses of a Small Group of Doctors, Boing Boing, Nov. 13, 2017, https://boingboing.net/2017/11/13/little-man-behind-the-curtain.html [https://perma.cc/4Q39-SMVJ].

154 See Mary Chris Jaklevic, MD Anderson Cancer Center's IBM Watson Project Fails, and so did the Journalism Related to it, Health News Rev. (Feb. 23, 2017), https://www.healthnewsreview.org/2017/02/md-anderson-cancer-centers-ibm-watson-project-fails-journalism-related/ (profiling MD Anderson Cancer Center's failure to implement WO) [https://perma.cc/62YQ-64JM].

155 Id.

156 Goldberg, Margo, The Robotic Arm Went Crazy! The problem of Establishing Liability in a Monopolized Field, 38 Rutgers Computer & Tech. L. J. 225, 249 (2012)Google Scholar (IBM could likely operate similarly to another company, Intuitive. “Because Intuitive has a monopoly, as well as limited expert witnesses, manipulation of an industry standard by such a dominant player is likely in a situation where they would be protecting themselves in a products liability lawsuit.”); Ross & Swetlitz, supra note 21 (“It is by design that there is not one independent, third-party study that examines whether Watson for Oncology can deliver. IBM has not exposed the product to critical review by outside scientists or conducted clinical trials to assess its effectiveness.”)

157 Ross & Swetlitz, supra note 21.

158 STAT - About Us, https://www.statnews.com/about/ (last visited March 25, 2018) (“We cast a critical eye on scientific discoveries, scrutinize corporate strategies, and chronicle roiling battles for talent, money, and market share. We examine controversies and puncture hype. With an award-winning newsroom, STAT gives you indispensable insights and exclusive stories on the technologies, personalities, power brokers, and political forces driving massive changes in the life science industry — and a revolution in human health.”) [https://perma.cc/49RC-JEHU].

159 Ross & Swetlitz, supra note 21.

160 Id.

161 Compare IBM Watson Health, https://www.ibm.com/watson/health/oncology-and-genomics/oncology/ (last visited October 18, 2017) (supporting WO's high performance and agreeance rates by showcasing and publishing a study) [https://perma.cc/3A63-3WUM] with Ross & Swetlitz, supra note 21 (“While it has emphatically marketed Watson for cancer care, IBM hasn't published any scientific papers demonstrating how the technology affects physicians and patients. As a result, its flaws are getting exposed on the front lines of care by doctors and researchers who say that the system, while promising in some respects, remains undeveloped.”)

162 Ross & Swetlitz, supra note 21.

163 Id. (“Doctors at Memorial Sloan Kettering acknowledged their influence on Watson. ‘We are not at all hesitant about inserting our bias, because I think our bias is based on the next best thing to prospective randomized trials, which is having a vast amount of experience,’ said Dr. Andrew Seidman, one of the hospital's lead trainers of Watson. ‘So it's a very unapologetic bias.’”)

164 Restatement (Third) of Torts § 15, cmt. a (2017) (“The basic rules governing causation in products liability litigation are the same as those governing tort law generally. It is unnecessary, therefore, to draft a set of causation standards limited in application to products liability; the prevailing rules and principles of causation may be applied in products litigation.”)

165 Mracek v. Bryn Mawr Hosp., 610 F.Supp.2d 401, 402 (E.D. Penn. 2009).

166 Id.

167 Mracek, supra note 166 at 404.

168 Id. at 405-07.

169 Norris v. Baxter Healthcare Corp., 397 F.3d 878, 881 (10th Cir. 2005) (distinguishing between general (actual) and specific (proximate) causation.)

170 Stephenson v. Upper Valley Family Care, Inc., 2010 WL 3610438 *4 (Ohio Ct. App. 2010) (“Stephenson presented several experts to testify regarding his expected life expectancy. Dr. Eliezer Nussbaum testified (via a transcript of his testimony from the first trial) that Stephenson's life expectancy had been decreased “by ten years or even more” due to Stephenson's lack of early treatment. Dr. Gary Mueller opined that Stephenson suffered a loss of life expectancy of between eight and twelve years due to delayed treatment.”)

171 Id. at *4-5.

172 Auto Owners Insurance Co. v. Seils, 871 N.W.2d 530, 545 (Mich. Ct. App. 2015).

173 Software is not within the scope of products liability. Childers, supra note 47 at 128 (2008).

174 Hardware and Software, http://cs.sru.edu/~mullins/cpsc100book/module02_introduction/module02-03_introduction.html (last visited Aug. 8, 2019) [https://perma.cc/FJB3-LRPE].

175 What is a CPU?, https://www.digitaltrends.com/computing/what-is-a-cpu/ (last visited Aug. 8, 2019) [https://perma.cc/AB3V-CHPC]; What is a Motherboard?, https://www.digitaltrends.com/computing/what-isa-motherboard/ (last visited Aug. 8, 2019) [https://perma.cc/2CKV-U2BT].

176 See id.

177 Ross & Swetlitz, supra note 21.

178 The standard by which the physician should be judged is the degree of knowledge and concomitant medical or surgical skill that a physician under the same or similar circumstances should “reasonably” possess. Pegalis, supra, note 89 at §§ 4:1.

179 I am referring to the fact that IBM has a monopoly on data related to WO. See supra note 157.

180 Dura Pharm., Inc. v. Broudo, 544 U.S. 336, 344 (2005).

181 Id. at 345.

182 Id. at 347.