The synergy and the interplay between the fields of Computer Science, Robots and AI has given to among others the regulatory and ethical issues[1]. There are two ways in which Technology and Law play together, one being the application of Technology in Law and the other being the regulation of Technology with Law. Both requires an understanding of how the Technology works and the Law regulates. The evolution of autonomous systems using machine learning to the most advanced robotics that has given rise to various applications like self-driving cars, drones, humanoids, search engines. Due to the autonomous nature of decision making at real time, there is a requirement for a more nuanced and a tailored approach for its regulation. The application of Artificial Intelligence has been pervasive through every industry including the Legal Industry. It is set to reduce the most mundane works and the inefficiencies across the Legal Industry. The use of Automation has been already benefited in the non-litigation especially due diligence of documents, digitizing case laws thereby allowing easier searches, storage, transfer etc., thereby decreasing the cost and labor hours drastically.

This research paper is split into majorly three parts apart from Introduction and Conclusion. In the first part, the possible applications of Artificial Intelligence in the Cross-Examination at the Courtroom probably decades into the future considering the present state of infrastructure in our Courtrooms is explored. In the second part, a scenario where a Face Lie Detector is used in the Courtroom and the possible legal issues that may arise from it is envisioned. In the third part, the new types of digital evidences that have arisen as a result of the development of Technology using Artificial Intelligence is analyzed with established Indian case laws. The illustrations of how these evidences help in determining the truth in the cases and also the admissibility of these evidences against the Constitutional provision of Self-Incrimination are also analyzed. The development of jurisprudence with respect to Artificial Intelligence and Digital Evidence has taken place primarily in the West and hence some of the scenarios are directly cited from foreign sources. In the end, to conclude the future of Artificial Intelligence, its limitations with respect to Law are also discussed in short.

1. Applications of Artificial Intelligence in Cross Examination

The Shadow of Artificial Intelligence has been cast across the fields, Legal Field is no exception. Every field has its share of monotony. Before digitization, Legal Field is known for its monotonous nature at the back of the office. Although AI has developed enough to automating document, helped in legal education, legal research, compliance etc., it is yet to breach into the walls of litigation especially in the lower judiciary. In this section, we will look at how Artificial Intelligence can aid in the entire process of Cross-Examination right from preparation to post Cross-Examination.

There are different subsets of Artificial Intelligence, the most relevant to the Legal Industry is Natural Language Processing and Speech Recognition as the Legal Field involves a lot of reading from documents and conversion of audio to text for further processing.

Every Cross-Examination requires a lot of legal research[2]. The efficiency of document analysis and the ability to flag a particular document as relevant for the research is increased by AI powered software that are available at market today. ROSS Intelligence is one such software powered by IBM and holds a major market share[3]. This allows the humans to concentrate on lesser number of specific documents rather identifying what is relevant and focus on tasks that require more human intervention like preparing questions for Cross-Examination. In civil cases, it can also help in performing due diligence that requires large number of analysis of documents.

Sentimental Analysis from Social Media[4] of Witnesses can also be done that could potentially change the outcome of the Cross-Examination. For example, if the witness gives an opinion that he is against death penalty and also states that he has always maintained the same. The opinion he had given on any of the Social media throughout the Internet could become relevant and will help considerably in countering the statement he makes in the Cross-Examination. Text Mining in Artificial Intelligence have been used by businesses throughout the world[5], it is only time that the application is also extended to Legal Field.

Information Technology equipped with Artificial Intelligence not only facilitates lawyers and judges as we saw in the preceding sections, it can also help in the administration of Courts. During and After the Cross-Examination, there requires a transcription of the process which can be done by the Speech Recognition tools that employ Artificial Intelligence. In Future, in cases the witness is from a different country speaking a different language, Artificial Intelligence could help the witness in the translation of the language of Court in real time[6]. Combing Speech-Recognition AI, Facial Recognition AI and Translation AI, it is now possible to read the lips with 95.6% accuracy at sentence level compared to 86.4% of the humans[7]. This can also help in analyzing video evidences, aiding the hearing impaired people in the Court procedure.

2. Legality of using Facial-Lie recognition in Courts

Machine Learning is used to detect deception or lies from non-verbal human behaviors reliably and with more efficiency than ever before. This uses micro expressions, movement of individual facial movements, body language with a system that is previously trained with a lot of data. Results have been proven to give accuracy to about 76.5%[8]. These AI Lie detectors have also been used in Border Crossing Points in the European Union[9]. These kind of technologies need not be restricted only to their intended use but should put to more broader applications potentially[10] including in Courts where it becomes even more prominent for faster adjudication and better evidentiary value without bias. DARE (Deception Analysis and Reasoning Engine), an AI System designed in University of Maryland and Dartmouth trained using the videos from the Courtroom[11]. DARE managed to spot 92% of the micro expressions. There is a significant use of Artificial Intelligence in this type of Lie Detector as model takes decisions based on the data fed to the system. The promising technology raises an important question whether it will be allowed in court for the purposes of Cross Examination and such.

Any introduction of new technology in the Court System has come by the initiation of the Courts itself. Example, in the recent judgment Swapnil Tripathi vs Supreme Court of India[12], the live telecast of the Supreme Court proceedings that are of constitutional importance having impact on public at large. This technology is analogous in legality to other such previous generation tests like Polygraph tests, Narco Analysis and Brain Mapping Tests. Although each tests are unique in its own way having respective drawbacks and advantages, but the fundamental objective of the test i.e. to determine the truth remains the same. Hence, the test applied to determine the evidentiary value, the constitutional validity will remain the same to the Facial Lie Detector using AI as well.

Self-Incrimination

In this light, one cannot ignore the precedent set by the case Selvi vs State of Karnataka[13], this case raises several issues pertaining to the self-incrimination, privacy and due diligence of the tests and has the majority opinion has been delivered by the Chief Justice.G. Balakrishnan himself. It was contended that the techniques like narco-analsis, lie detector and brain mapping violated an accused person’s right against self-incrimination under Article 20(3) of the Constitution and their right to life and personal liberty under Article 21 of the Consitution. This Court referred to the case M.P. Sharma v Satish Chandra[14] where the issue was whether search and seizure of documents would amount to Self-Incrimination. The Court held that the words “to be witness” included self-incrimination through documentary evidence, as it included any positive volitional act that furnishes evidence. Comparing it with Facial Lie Detector that can be used in Courts, it is not self-incriminating as there is no question of volition here. The Court takes into discretion to analyze the accused his behavioral changes. The system only helps or augments the ability of the Judge to see things he couldn’t previously see including his micro facial expressions, his analysis of language etc. There is no compulsion on the accused too. Also, the technology used does not become part of the investigation process but it forms part of the Cross-Examination i.e. in the presence of the judiciary where the witness is free to testify anything not necessarily against himself.
Further in support, in the Kathi Kalu case[15], realizing that in MP Sharma Case would exclude some important criminal procedures, and would invalidate some of the useful provisions in the Indian Evidence Act,1872. The Court therefore interpreted further deeper the phrase “a positive volitional act” and equated “volition” with the ability to alter a state of affairs like giving a person’s fingerprint or handwriting. Such act only furnish evidence and does not amount “to being a witness”. Similarly, the AI Lie Detector analyzing the witness does not change anything and hence will not amount to being a witness in case if the witness is accused.

Evidentiary Value

However, using the results cannot be taken as a conclusive proof but can only be considered as a corroborative evidence as any Artificial Intelligence System depends on the data it is being fed. Also, there is a possibility that the witness gives statements in a half-heartedly manner not sure of its authenticity. At such times, it is probable that the systems might infer it to be false. If it is an expert system, where it predicts based upon the past data that is fed, the output depends on the data that is being fed[16]. In non-expert systems, hard programming errors can lead to inaccurate results. Though this technology cannot be used as an evidence, it can be used to see things which a judge would miss with a normal sight. There is no technology that is completely fail proof. Our criminal system is built upon “beyond reasonable doubt” and not on “beyond all doubts”. Adopting the technology and relying on it the right amount with the judges understanding the technology so that it does not form any bias will help the judiciary in the long run. The necessary guidelines has to be set-up by the judiciary like in the case of lie-detector tests for investigation purpose.[17]

3. Digital Evidence in the world of AI and Internet of Things

Internet of Things is the network of physical objects that have been connected to Internet thereby transferring, collecting and exchanging data for the functioning[18]. They usually contain sensors that captures data and use Artificial Intelligence with the data available and take the next step. For example, based on the previous data and artificial intelligence, a self-driving car calculates the most optimal route to reach a destination. The number of people worldwide was surpassed by the number of “things” connected to the Internet at 2009[19] and this was only considered as the beginning of the IoT movement[20]. It is estimated that by 2020, there will be up to 50 billion connected devices. The basic premise is that these objects communicate with embedded sensors and wired/wireless communication protocols via apps, browsers and apps. The legal community need to take advantage of the new revolution in the digital evidence that could potentially change their practice in the courtroom. These key evidences from the IoT devices somewhere hidden in the internet can make cases clearer and present a story than ever before. These evidences might be right in front of our faces.

It can make and add to the credibility or discredit entirely a witness’s statement. For example, data that is obtained from a Fitness band has been used as an evidence to show a person’s diminished physical activity resulting in an injury during work time thereby claiming compensation[21]. Wearable devices collect data and use algorithms to analyze data and report trends compared to the general population. IoT with Artificial Intelligence has also been majorly developed in transportation for the creation of self-driving vehicles. These smart vehicles record real time navigation, person inside the vehicles, biometric data, state of the vehicle, surroundings and environment, weather etc.[22] In a road-accident case or a claim to insurance, all these data accumulated by the vehicle can be used as evidence to determine the facts.

IoT Object as Witness

As more and more of IoT Devices become relevant as witness and evidence, we need to figure out the weightage such evidence will bear. If there is a contradiction between a human witness and an electronic evidence, which will carry more weightage and be given preference by the judge. One cannot be sure that electronic evidence is always right, both the systems are fallible although normally the systems should be credible as they do not intentions, bias etc. unless tampered or extra ordinary circumstances exist to disprove the evidence. In order to give weightage to the evidence, the judge must be able to understand the working of that device. Experts can be appointed to determine the credibility in such circumstances. Courts will also need to take care of Article 20(3) of the Constitution wherever needed although as previously discussed these evidence must be kept out of the ambit of Self Incrimination. The more of such cases happen, people are more likely to refrain from using such devices. Perhaps, in a few decades, instead of Eye Witness we will have Google Glass acting as an evidence.

Primarily it shall be the duty of the lawyer to get right evidence, understanding the technology and convincing the judge for the admissibility of the evidence and such a scenario will be inevitable in the Information Age.

Admissibility of New Age Evidences

There is no dispute regarding the admissibility of electronic evidence but the contention always lies, on the extent to which it is reliable. In the case of Shafhi Mohammad and Ors. v. The State of Himachal Pradesh and Ors.[23], it was held that the investigating agencies are not fully equipped or prepared for the use of videography, but the time is ripe to take initiation and introduction of videography in the investigation process[24]. In the judgement it has also referred to other judgements like Ram Singh and Ors v. Col. Ram Singh[25], English judgements[26] and American Jurisprudence[27] where it was approved to the effect that “it will be wrong to deny to the law of evidence advantages to be gained by new techniques and new devices, provided the accuracy of the recording can be proved. The only difference between the evidence spoken in the above judgement and evidence obtained from the IoT is that it is already present unlike the evidence created using the videography during the process of investigation.

The issues arise how these data will be submitted to the Court, whether they will be considered as Primary Evidence or Secondary Evidence. What will be the compliance requirement for such evidences is to explored. These evidences obtained from devices can be compared to the call data records obtained from the mobile phone operator. In the State (N. C. T. of Delhi) vs Navjot Sandhu[28], the infamous Parliamentary attack case, the accused raised the issue of inadmissibility of electronic records i.e. the mobile data records as the credibility and reliance of the telephone records are questionable because the prosecution had not complied with sub-section (2) of the Section 65B of the Indian Evidence Act. But, the Court rejected the argument and concluded that the cross-examination of competent witness acquainted with the functioning of the computer and the manner in which the print out of the call records were taken were sufficient to prove the call records.

Most of the electronic evidences are submitted as secondary evidences, as producing primary evidence becomes nearly impossible with respect to the definition of the primary evidence in the statute[29]. The original data or the “contents of the documents” according to the statute are situated in the servers of remote location which makes it an obligation to make a copy of the record and hence the production of certificate. In previous Supreme Court judgements to Anvar vs Basheer[30], the use of the word “may” as interpreted does not preclude the parties from adducing electronic records using other traditional Sections 63 and 65[31]. Many High Court judgements followed suit and Rakesh Kumar vs State[32] is an example of this. But after the Anvar P.V v P.K Basheer[33] case, it was overturned. The Supreme Court mandated for the compliance of Section 65B of the Indian Evidence Act,1872.

This is actually good in terms of next gen evidences as doing so will require a certificate from an expert. This makes sure that the evidence was not tampered. Computer Evidences are susceptible to tampering. But Computer Forensic Experts is continually developing in parallel to find if the evidence was really tampered and if so in what manner. Access, Obtaining and Securing Evidence will also be relevant to Data Protection which does not form part of the scope of this paper.

Limitations and Conclusion:
The future and capabilities of Artificial Intelligence is uncertain[34] and the fact that Artificial Intelligence can learn over time and data, means that it is only bound to improve in the future. The recent potential of Artificial Intelligence has been realized only because of the growth in the volume of data. When the data in the legal industry is structured, there is no stopping of automating mundane work and tasks that require less human intelligence that is prevalent in the industry. Engineers have been able to create general purpose AI systems that mimics human brain using machine learning and neural network processing[35].

While human-like deductive reasoning, inference, and decision-making by a computer is still a long time away, there have been remarkable gains in the application of AI techniques and associated algorithms[36]. There are also other significant limitations in implementing full-fledged Artificial Intelligence in the Court system, as the legal ethics of what is being fed to the system through data and algorithms are always questionable. In the Cross Examination and Courtroom, several other factors play as in the current set up, there is lot of disturbance, multiple conversations at interplay, the supremacy of the judge, the empathy of the jury etc.

We as Indians, have always been a decade or two behind the West in the adoption of technology. The Legal Field has been no different. Some of the scenarios discussed in this paper is still yet to be tested across the World and there is no prior precedent like the use Facial Lie Detector, using Sentimental Analysis of the witness to contradict their statements. We are still a long way from implementing these in our courts as the digital revolution is fairly new in our society. But as we have already entered the digital age, it is evitable to not to adapt to changes it has brought with it. The creation and usage of new tools are never going to stop. To conclude, Artificial Intelligence is here to stay to augment the human beings to do their jobs better and concentrate on higher level tasks like negotiation, representing clients in Courts which will make the access to justice to the remotest person of the society at a much lower cost. It is not a question of a Machine or a Human but it is the Machines and the Humans in the future.

End-Notes
[1] Pagallo, U. (2011). Robots of just war: A legal perspective. Philosophy and Technology, 24(3), 307–323.
[2] Bernard Marr, How AI and Machine Learning Are Transforming Law Firms and The Legal Sector, https://www.forbes.com/sites/bernardmarr/2018/05/23/how-ai-and-machine-learning-are-transforming-law-firms-and-the-legal-sector/#51e1095632c3
[3] https://rossintelligence.com/
[4] Elliott Asha, Daniel L. Chenb, Sergio Gallettac, Judicial Sentiments and Social Attitudes: Evidence from U.S. Circuit Courts, https://users.nber.org/~dlchen/papers/Judicial_Sentiments_and_Social_Attitudes.pdf
[5] Shashank Gupta, Applications of Sentiment Analysis in Business, https://towardsdatascience.com/applications-of-sentiment-analysis-in-business-b7e660e3de69
[6] Yogesh Singh, Using AI and Deep Learning for Transcription and Translation, https://medium.com/@cyogesh56/using-ai-and-deep-learning-for-transcription-and-translation-d7a44176336f
[7] LipNet: End-to-End Sentence-level Lipreading, Yannis M. Assael, Brendan Shillingford, Shimon Whiteson, Nando de Freitas, https://arxiv.org/abs/1611.01599
[8] Dr. Mahbub Alam Majumdar, Using Machine Learning for Lie Detection: Classification of Human Visual Morphology, http://dspace.bracu.ac.bd/xmlui/bitstream/handle/10361/10144/14101005%2C17241023%2C17241022_CSE.pdf?sequence=1&isAllowed=y
[9] Dani Deahl, The EU plans to test an AI lie detector at border points https://www.theverge.com/2018/10/31/18049906/eu-artificial-intelligence-ai-lie-detector-border-points-immigration
[10] Jeff Daniels, Lie-detecting computer kiosks equipped with artificial intelligence look like the future of border security, https://www.cnbc.com/2018/05/15/lie-detectors-with-artificial-intelligence-are-future-of-border-security.html
[11] The robot that knows when you're lying: Scientists create an AI that can detect deception in the courtroom (and it's already 'significantly better' than humans) , https://www.dailymail.co.uk/sciencetech/article-5197747/AI-detects-expressions-tell-people-lie-court.html
[12] Swapnil Tripathi vs Supreme Court of India, WRIT PETITION (CIVIL) NO. 1232 OF 2017, https://www.sci.gov.in/supremecourt/2017/40426/40426_2017_Judgement_26-Sep-2018.pdf
[13] Smt Selvi & Ors. V State of Karnataka, 2010 7 SCC 263
[14] M. P. Sharma And Others v Satish Chandra, 1954 SCR 1077
[15] State of Bombay v Kathi Kalu Oghad, 1962 3 SCR 10
[16] Expert system, https://en.wikipedia.org/wiki/Expert_system
[17] Guidelines on Administration of Lie Detector Test, http://nhrc.nic.in/press-release/guidelines-administration-lie-detector-test
[18] Embedded Intelligence – Connecting Billions of Smart Sensors into the Internet of Things, ARM Holdings, http://ir.arm.com/phoenix.zhtml?c=197211&p=irol-embeddedintelligence, archived at https://perma.cc/3HWX-QBWW (last visited Mar. 23, 2016).
[19] Dave Evans, Cisco Internet Bus. Solutions Grp., The Internet of Things: How the Next Evolution of the Internet Is Changing Everything 3 (2011), http://www.cisco.com/c/dam/en_us/about/ac79/docs/innov/IoT_IBSG_0411FINAL.pdf, archived at https://perma.cc/HDF9-NM6T.
[20] Accenture, The Internet of Things: The Future of Consumer Adoption (2014), https://www.accenture.com/t20150624T211456__w__/us-en/_acnmedia/Accenture/Conversion-Assets/DotCom/Documents/Global/PDF/Technology_9/Accenture-Internet-Things.pdf, archived at https://perma.cc/JKG7-UT4P.
[21] Kate Crawford, When Fitbit is the Expert Witness, The Atlantic (Nov. 19, 2014), http://www.theatlantic.com/technology/archive/2014/11/when-fitbit-is-the-expert-witness/382936/, archived at https://perma.cc/AW5G-5NY2.
[22] An EDR is “a device or function in a vehicle that records the vehicle’s dynamic time-series data during the time period just prior to a crash event (e.g., vehicle speed vs. time) or during a crash event . . . intended for retrieval after the crash event.” 49 C.F.R. § 563.5 (2015). Telematics refers to data collection transmission, and processing technologies for use in vehicles.
[23] Shafhi Mohammad and Ors. v. The State of Himachal Pradesh and Ors, Special Leave Petition (Criminal) Nos. 2302 of 2017
[24] Para 10, ibid 17
[25] Ram Singh & Ors vs Col. Ram Slngh, 1985 SCC 611
[26] R. v. Maqsud Ali,(1965) 2 All ER 464, and R. v. Robson, (1972) 2 ALL ER 699
[27] American Jurisprudence 2d (Vol 29) Page 494
[28] State (N.C.T. Of Delhi) vs Navjot Sandhu, 2005 CrLJ 3950
[29] The Indian Evidence Act, 1872
[30] Anvar P.V. v. P.K. Basheer and Ors., AIR 2015 SC 180.
[31] Indian Evidence Act,1872
[32] Rakesh Kumar v. State & Ors., 2009 DLT 658.
[33] Supra 23
[34] Artificial Intelligence and its role in Near Future, https://arxiv.org/pdf/1804.01396.pdf
[35]Computer That Can Closely Mimic Human Brain’s Neural Network, https://www.evolving-science.com/intelligent-machines/computer-mimic-human-brains-00721
[36] Artificial Intelligence, Deep Learning, and Neural Networks Explained, https://www.innoarchitech.com/artificial-intelligence-deep-learning-neural-networks-explained/