A better understanding of AI starts with its definition. What constitutes AI varies considerably depending on whom you ask. According to Merriam-Webster’s broad characterization, AI is “the capability of a machine to imitate intelligent human behavior.”3 Similarly, John McCarthy, the “father” of AI, defined it as “the science and engineering of making intelligent machines, especially intelligent computer programs. It is related to the similar task of using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable.”4 Tech giant Amazon, Inc., which relies on AI for a host of tasks, defines it as “the field of computer science dedicated to solving cognitive problems commonly associated with human intelligence, such as learning, problem solving, and pattern recognition.”5 A different perspective focuses on differences between “soft” and “hard” AI, noting the former referred to menial tasks performed by computers while the latter represented the ability to perform complex reasoning so often displayed on the big screen.6 Another source defined AI simply as “computers that think.”7
Using a broad definition, AI is currently being employed in exceptionally useful ways in a variety of settings. For example, AI is being embraced by a number of agencies within the U.S. government. The Social Security Administration is using AI and algorithms “to speed sorting and processing.”8 The Environmental Protection Agency is deploying AI to better identify and locate factories that should be inspected for illegal polluting.9 The Department of Transportation is currently “researching, developing, and testing automatic traffic accident detection based on video” to increase and maintain commuter traffic and efficiency.10 Researchers continue to look for other ways U.S. agencies can improve efficiencies by employing AI.11 AI has similarly thrived in the U.S. military, with exponential potential for future growth.12 AI has become thoroughly engrained within the U.S. government and continues to provide benefits that would otherwise be unobtainable.13
The private sector has perhaps been even more zealous to adopt and implement AI. Netflix, Amazon, Google, Apple, Tesla, and Pandora all reap substantial benefits from AI.14 Each of these companies uses AI in part to learn from past consumer behavior and predict what consumers will want in the future.15 Moreover, AI can improve customer satisfaction with a company while also reducing costs.16 AI can make customer interactions more efficient and effective by addressing, rerouting, or escalating customer inquests to send them to a human operator.17 AI also can help with more efficient and effective fraud detection. For example, PayPal uses “volumes of data to continuously train their fraud detection algorithms to predict and recognize” fraud.18 Investment in AI is forecasted to continue at a rapid pace.19 By 2023, worldwide spending on AI could reach $98 billion.20 AI has created numerous efficiencies in society and is poised to bring about further substantive change.
AI is also being deployed in the field of law. According to some researchers, attorneys at big law firms complete only 4 percent of all document review, with the rest done by AI.21 Indeed, at least one “artificial intelligence lawyer,” apparently related to IBM’s Jeopardy-winning Watson program, has been “hired” by a law firm.22 Judges and lawyers also currently use AI when analyzing appropriate sentences under the Federal Sentencing Guidelines.23 On the criminal side, police departments are turning to AI in an attempt to reduce bias and predict crimes.24 Moreover, AI is finding applications in forensic science, such as analyzing DNA and interpreting radiological images to assist coroners in establishing a cause of death.25 And a court in England recently found that law enforcement properly used automatic facial recognition software during a criminal investigation.26
Other law-related AI applications are more case-specific. A recent AI mediator in Canada helped two parties settle a claim after a human mediator could not close the deal.27 Another case involved a disputed claim over ownership of a homerun baseball between two fans.28 A researcher converted human arguments into computer code, fed the code into an AI program, and obtained results similar to those found by the California Supreme Court.29
Similarly, the future development and use of AI have immeasurable potential to support the judicial system. Efficiently researching case law has progressed leaps and bounds in the last 25 years. Prospective AI programs could make another huge leap by reducing hours of case law research into seconds.30 Moreover, AI programs could lead to consistently fairer outcomes, thereby increasing the public’s confidence in the judicial system while simultaneously reducing crime rates and jail populations.31 Such programs also could be used to help identify and remedy previous decisions that were unjust.32 And AI programs could increase access to justice, especially for those with less means.33 These benefits would assist the judiciary with both menial and substantive tasks while also improving the overall system.
Interestingly, at least some of these benefits are closer to fact than science fiction. A pilot judicial AI program in Estonia is set to begin adjudicating claims where less than roughly $8,000 is in dispute.34 Although still in its early stages, the program will likely focus on contractual disputes.35 Crucially, decisions will be appealable to a human judge.36 In the same vein, AI developed overseas correctly predicted the outcome of 79 percent of the cases it analyzed that were before the European Court of Human Rights.37 In South Korea, legal experts will soon engage in a competition with AI to review various contracts.38 Indeed, perhaps in recognition of these innovations, the European Union recently enacted ethical guidelines for companies working to develop and implement AI.39 It seems like a small bridge to cross before such programs can be implemented to assist judges and the judicial system to reach the best result.40 AI continues to develop at a breakneck pace, bringing with it the potential for numerous benefits within the judiciary. While some speculate AI could replace lawyers and judges altogether, it appears more likely that technology and AI will work to increase the efficiency and effectiveness of the legal profession.41 New technology will instigate change, and lawyers and judges will continue to adapt.
Despite its benefits, the implementation of AI has brought and continues to create the potential for glitches. Jeff Ward,42 writing for Judicature, recently outlined several helpful points all judges should know when interacting with AI.43 For one, treating AI like incomprehensible “magic” could lead to issues for the judicial system. It is important to realize that “all AI tools are the product of intentional human design, woven from long-held understandings of statistical methodologies, the growing insights of computer science, and the human mind.”44 Approaching AI as baffling could “inadvertently cede responsibility and agency to tech companies and limit the roles that other stakeholders can play.”45 Judges must vigilantly avoid the urge to sweep the inputs and mechanics behind AI away as inexplicable.
Likewise, AI’s complexity may implicate transparency issues in important areas.46 For example, AI’s use in the formulation of a presentence investigation report led to due process concerns in at least one case.47 Specifically, one issue concerned an algorithm’s assignment of weight to an individual’s gender and criminogenic factors.48 The court ultimately found no violation of due process.49 Another case challenged AI’s use in determining aggravated factors in sentencing decisions.50 Such cases highlight the importance of transparency and understanding involved with AI.51 Complicating this issue is the fact that AI is designed to learn, grow, and evolve within established parameters.52 The judicial system should be keenly aware of such parameters to ensure AI does not develop to conflict with any relevant laws or rules.
Next, judges should engage, engage, engage with technology.53 Simply because the practice of law has stayed substantially the same for the last few decades does not mean it will continue to do so for the next 10 years. Technology has rapidly evolved this century and will likely continue to do so at an explosive pace.54 For instance, pretrial discovery and in-court presentations have transformed from paper and poster boards to digital data and HD monitors.55 Even handwritten letters have become a societal casualty to email, texts, and social media.56 The judicial system will have an easier time adapting to future changes that could be transformative if it stays abreast of recent technological innovations. AI could be implemented much more easily in the future with well-informed, up-to-date judges.
In the same vein, judges should be wary of potential fakes. AI programs continue to grow more powerful and sophisticated. Such progress brings with it the ability to create and spread “hyper-realistic creations that are significant both in and out of the courtroom.”57 In other words, there now exist relatively accessible ways to create fake videos, photos, and voices, among other things, that are thoroughly convincing.58 In fact, an app in China recently exploded in popularity after allowing users to swap their face with a character from famous movie clips.59 Users can now see themselves act out 30-second clips of Titanic or Game of Thrones with a few taps of a touch screen.60 Likewise, fraudsters were able to steal $243,000 by using AI to mimic the voice of a company’s CEO.61 The thieves appear to have perfectly mimicked the CEO’s voice, tone, punctuation, and German accent.62 The ease of use and access of such technology have broad implications for numerous sectors, particularly with the courts. Judges who stay informed and engaged with technological progress will be in the best position to adapt to new hazards such as fake digital media content.
While perhaps not as dramatic as a gun-toting terminator or smooth-talking computer, artificial intelligence has nonetheless arrived. AI has impacted and will continue to impact the judicial system. Portions of the U.S. government, as well as many companies operating in the U.S. economy and abroad, have extensively implemented AI to the benefit of society and the bottom line. Rather than shying away from change, the judiciary’s best way forward is to engage and understand these new technological innovations. It is imperative for judges who interact with AI to know the basics of how it works, including its uses, data inputs, and limitations. Artificial intelligence will likely only grow in its contributions to the world. Judges must be sure to understand and realize these innovations for the benefit of all.
Endnotes
1. Alan Kay is a famous computer scientist. Alan Kay, American Computer Scientist, Britannica, https://www.britannica.com/biography/Alan-Kay (last visited Sept. 11, 2019).
2. However, such a future may not be as farfetched as it seems. Corey S. Powell, Cyborgs Will Replace Humans and Remake the World, James Lovelock Says, NBC News (Aug. 29, 2019, 9:02 AM), https://www.nbcnews.com/mach/science/cyborgs-will-replace-humans-remake-world-james-lovelock-says-ncna1041616.
3. Artificial Intelligence, Merriam-Webster, https://www.merriam-webster.com/dictionary/artificial%20intelligence (last visited Sept. 11, 2019).
4. William J. Connell, Artificial Intelligence in the Legal Profession—What You Might Want to Know, 66 R.I.B.J., no. 6, May/June 2018, at 5.
5. Bernard Marr, The Key Definitions of Artificial Intelligence (AI) That Explain Its Importance, Forbes (Feb. 14, 2018, 1:27 AM), https://www.forbes.com/sites/bernardmarr/2018/02/14/the-key-definitions-of-artificial-intelligence-ai-that-explain-its-importance/#1ffd99274f5d.
6. David E. Chamberlain, Artificial Intelligence and the Practice of Law or Can a Computer Think Like a Lawyer?, St. B. Tex., Sept. 2016, at 1, www.texasbarcle.com/cle/OLViewArticle.asp?a=184155&t=PDF&e=14730&p=1.
7. Connell, supra note 4.
8. Eric Niler, Can AI Be a Fair Judge in Court? Estonia Thinks So, Wired (Mar. 25, 2019, 7:00 AM), https://www.wired.com/story/can-ai-be-fair-judge-court-estonia-thinks-so.
9. Id.
10. Christopher Rigano, Using Artificial Intelligence to Address Criminal Justice Needs, NIJ J. no. 280, Jan. 2019, at 36, 38, https://www.ncjrs.gov/pdffiles1/nij/252031.pdf.
11. Niler, supra note 8.
12. Zachary Fryer-Biggs, Coming Soon to a Battlefield: Robots That Can Kill, Atlantic (Sept. 3, 2019, 5:00 AM), https://www.theatlantic.com/technology/archive/2019/09/killer-robots-and-new-era-machine-driven-warfare/597130; Samuel Gibbs, Google’s AI Is Being Used by US Military Drone Programme, Guardian (Mar. 7, 2018), https://www.theguardian.com/technology/2018/mar/07/google-ai-us-department-of-defense-military-drone-project-maven-tensorflow.
13. Somewhat similarly, AI could likely benefit developing countries by increasing yields from urban farming. Daphne Ewing-Chow, Combining Artificial Intelligence with Urban Farming Can Be a Game Changer for Developing Countries, Forbes (Sept. 1, 2019, 12:19 AM), https://www.forbes.com/sites/daphneewingchow/2019/09/01/combining-artificial-intelligence-with-urban-farming-can-be-a-game-changer-for-developing-countries/#27042d67d4b2.
14. R.L. Adams, 10 Powerful Examples of Artificial Intelligence in Use Today, Forbes (Jan. 10, 2017, 8:32 AM), https://www.forbes.com/sites/robertadams/2017/01/10/10-powerful-examples-of-artificial-intelligence-in-use-today/#70918c72420d.
15. Id. For example, Netflix collects data from a user’s historical movie or television show choices to offer unique and personalized recommendations on which shows to watch next. Mario Gavira, How Netflix Uses AI and Data to Conquer the World, linkedin (July 2, 2018), https://www.linkedin.com/pulse/how-netflix-uses-ai-data-conquer-world-mario-gavira. A reported 75 percent of users pick shows based on Netflix’s recommendations. Id.
16. Dawson Whitfield, What Happens to Designers When the Benefits of Artificial Intelligence Outweigh the Costs?, Forbes (Sept. 4, 2019, 7:45 AM), https://www.forbes.com/sites/forbestechcouncil/2019/09/04/what-happens-to-designers-when-the-benefits-of-artificial-intelligence-outweigh-the-costs/#712a4003a121; Bernard Marr, The Amazing Ways Telecom Companies Use Artificial Intelligence and Machine Learning, Forbes (Sept. 2, 2019, 12:22 AM), https://www.forbes.com/sites/bernardmarr/2019/09/02/the-amazing-ways-telecom-companies-use-artificial-intelligence-and-machine-learning/#2bbcd5884cf6.
17. Marr, supra note 16.
18. Rigano, supra note 10. Specifically, Paypal’s fraud analysts examine real-world situations that have triggered fraud warnings in the past. Penny Crosman, How PayPal Is Taking a Chance on AI to Fight Fraud, Am. Banker (Sept. 1, 2016), https://www.americanbanker.com/news/how-paypal-is-taking-a-chance-on-ai-to-fight-fraud. Analysts then “develop scenarios for good and bad user behavior that they . . . feed into the [AI] program,” thereby making the program more effective at detecting fraud. Id.
19. Press release, Int’l Data Corp., Worldwide Spending on Artificial Intelligence Systems Will Be Nearly $98 Billion in 2023, According to New IDC Spending Guide (Sept. 4, 2019), https://www.idc.com/getdoc.jsp?containerId=prUS45481219.
20. Id.
21. Connell, supra note 4.
22. Id.
23. Current AI Helps Judges and Lawyers When Analyzing Appropriate Sentences Under Federal Sentences Guidelines, 44 Am. Jur. Trials 1, § 14 (originally published in 1992).
24. Katie Brigham, Courts and Police Departments Are Turning to AI to Reduce Bias, but Some Argue It’ll Make the Problem Worse, CNBC (Mar. 17, 2019, 9:30 AM), https://www.cnbc.com/2019/03/16/artificial-intelligence-algorithms-in-the-criminal-justice-system.html.
25. Rigano, supra note 10.
26. Jenny Rees, South Wales Police Use of Facial Recognition Ruled Lawful, BBC (Sept. 4, 2019), https://www.bbc.com/news/uk-wales-49565287.
27. Tara Vasdani, From Estonia AI Judges to Robot Mediators in Canada, U.K., Law. Daily (June 13, 2019, 11:47 AM), https://www.thelawyersdaily.ca/articles/12997.
28. Larry N. Zimmerman, Artificial Intelligence in the Judiciary, 85 J. Kan. B. Ass’n, no. 7, July/Aug. 2016, at 20, 20.
29. Id.
30. See Connell, supra note 4 at 5, 7.
31. Arthur Rizer & Caleb Watney, Artificial Intelligence Can Make Our Jail System More Efficient, Equitable, and Just, 23 Tex. Rev. L. & Pol. 181, 226–27 (2018).
32. Angela Chen, How Artificial Intelligence Can Help Us Make Judges Less Biased, Verge (Jan. 17, 2019, 12:07 PM), https://www.theverge.com/2019/1/17/18186674/daniel-chen-machine-learning-rule-of-law-economics-psychology-judicial-system-policy.
33. Joel Tito, How AI Can Improve Access to Justice, Ctr. Pub. Impact (Oct. 23, 2017), https://www.centreforpublicimpact.org/joel-tito-ai-justice.
34. Niler, supra note 8.
35. Id.
36. Id.
37. Andrew Griffin, Robot Judges Could Soon Be Helping with Court Cases, Independent (Oct. 24, 2016, 10:06 AM), https://www.independent.co.uk/life-style/gadgets-and-tech/news/ai-judge-robot-european-court-of-human-rights-law-verdicts-artificial-intelligence-a7377351.html.
38. H.M. Kang, Artificial Intelligence to Compete with Legal Experts on Legal Review, Kor. Bizwire (Aug. 8, 2019), http://koreabizwire.com/artificial-intelligence-to-compete-with-legal-experts-on-legal-review/142022.
39. Foo Yun Chee, AI Must Be Accountable, EU Says as It Sets Ethical Guidelines, Reuters (Apr. 8, 2019, 7:02 AM), https://www.reuters.com/article/us-eu-ai/ai-must-be-accountable-says-eu-as-sets-ethical-guidelines-idUSKCN1RK1CV. Similarly, the Pentagon is currently seeking an “ethicist” to oversee military AI. David Smith, Pentagon Seeks “Ethicist” to Oversee Military Artificial Intelligence, Guardian (Sept. 7, 2019, 2:00 PM), https://www.theguardian.com/us-news/2019/sep/07/pentagon-military-artificial-intelligence-ethicist.
40. Interestingly, an AI program recently aced an eighth-grade science quiz. Cade Metz, A Breakthrough for A.I. Technology: Passing an 8th-Grade Science Test, N.Y. Times (Sept. 4, 2019), https://www.nytimes.com/2019/09/04/technology/artificial-intelligence-aristo-passed-test.html?smid=nytcore-ios-share. Such a feat required making connections using logic, and not simply retrieving facts stored in memory. Id. The same program earned a “B” on a 12th-grade science quiz. Id. AI appears poised to better assist judges using the ability to reason logically.
41. Predicting judges and lawyers will become “less necessary” as AI continues to evolve. Daniel Ben-Ari et al., “Danger, Will Robinson”? Artificial Intelligence in the Practice of Law: An Analysis and Proof of Concept Experiment, 23 Rich. J.L. & Tech. 3, 57–63 (2017).
42. Jeff Ward is the Associate Dean of Technology and Innovation and serves as the director of Duke University’s Center on Law & Technology (DCLT). DCLT coordinates Duke’s leadership at the intersection of law and technology with programs such as the Duke Law Tech Lab, a pre-accelerator for legal technology companies. DCLT also assists with the Access Tech Tools Initiative, a program to help students and Duke’s community partners employ human-centered design thinking and available technologies to create tools to enhance access to legal services.
43. Jeff Ward, 10 Things Judges Should Know About AI, 103 Judicature, no. 1, Spring 2019, at 12, https://judicialstudies.duke.edu/2019/04/10-things-judges-should-know-about-ai.
44. Id. at 13.
45. Id.
46. For a discussion on potentially discriminatory AI, see Ignacio N. Cofone, Algorithmic Discrimination Is an Information Problem, 70 Hastings L.J. 1389 (2019).
47. Ward, supra note 43; State v. Loomis, 881 N.W.2d 749 (Wis. 2016).
48. Ward, supra note 43; Loomis, 881 N.W.2d at 764–68.
49. Ward, supra note 43; Loomis, 881 N.W.2d at 764–68, 772.
50. Ward, supra note 43; Malenchik v. State, 928 N.E.2d 564, 573 (Ind. 2010). Additional discussion and analysis of these cases is available at John Lightbourne, Damned Lies & Criminal Sentencing Using Evidence-Based Tools, 15 Duke L. & Tech. Rev. 327 (2017).
51. Ward, supra note 43.
52. Kalev Leetaru, AI Correctness Is Not the Same as AI Ethics, Forbes (Aug. 29, 2019, 5:59 PM), https://www.forbes.com/sites/kalevleetaru/2019/08/29/ai-correctness-is-not-the-same-as-ai-ethics.
53. Ward, supra note 43.
54. The Honorable Herbert B. Dixon Jr., Technology Changes Coming Faster and Faster, Judges’ J., Fall 2014.
55. Id.
56. Id.
57. Ward, supra note 43.
58. Id.
59. David Ingram, A Face-Swapping App Takes off in China, Making AI-Powered Deepfakes for Everyone, NBC News (Sept. 5, 2019, 6:39 AM), https://www.nbcnews.com/tech/security/face-swapping-app-takes-china-making-ai-powered-deepfakes-everyone-n1049501.
60. Id.
61. Edward Ongweso Jr., Thieves Used Audio Deepfake of a CEO to Steal $243,000, Vice (Sept. 5, 2019, 12:23 PM), https://www.vice.com/en_us/article/d3a7qa/thieves-used-audio-deep-fake-of-a-ceo-to-steal-dollar243000.
62. Id.