Justice Revolutionized - The Promise And Perils Of Artificial Intelligence On The Criminal Justice System
Author: LawSphereX
Category: Criminal Law
Published:
Updated:
JUSTICE REVOLUTIONIZED: THE PROMISE AND PERILS
OF ARTIFICIAL INTELLIGENCE ON THE CRIMINAL JUSTICE SYSTEM Krishna Rao Alaparthi, LL.M (Criminology & Criminal Justice System), LL.M (Corporate and Securities Laws), Osmania University, Hyderabad. Samraji Himabindu, LL.M (Corporate and Securities Laws), P. G. Diploma in Cyber Laws (NALSAR)., Osmania University, Hyderabad.
ABSTRACT
Artificial Intelligence (AI) is not just a technological marvel; it is rapidly becoming a pivotal force in reshaping the criminal justice system. From predicting crimes before they happen to influencing the fate of individuals in courtrooms, AI’s integration promises a revolution in how justice is administered. On one hand, AI tools like predictive policing algorithms and risk assessment systems have the potential to reduce crime and ensure more informed judicial decisions. These tools can analyze vast amounts of data to identify patterns and trends that humans might miss, leading to more effective crime prevention strategies and fairer sentencing. Imagine a world where law enforcement agencies can anticipate criminal activities and intervene before they occur, or where judges can make decisions based on comprehensive data analysis rather than subjective judgment.
On the other hand, these same tools risk perpetuating existing biases, lack transparency, and raise serious concerns about accountability and civil liberties. AI systems are only as good as the data they are trained on, and if that data reflects societal biases, the AI will too. This could lead to discriminatory practices and unjust outcomes. Moreover, the opaque nature of AI algorithms undermine trust on judiciary. This paper explores the profound impact of AI on various organs of the Criminal Justice System like Police, Prosecution, and Judiciary. Also, highlights six essential principles of responsible AI (Fairness, Reliability & safety, Privacy & security, Inclusiveness, Transparency & Accountability).
INTRODUCTION
The Criminal Justice System, a cornerstone of modern society, exists to uphold laws, protect citizens, and ensure fairness. Throughout history, the system has evolved in response to cultural, legal, and technological advancements. Today, we stand at the threshold of another revolution, one driven by Artificial Intelligence (AI). Artificial Intelligence (AI) is not just a technological marvel; it is rapidly becoming a pivotal force in reshaping the criminal justice system. From predicting crimes before they happen to influencing the fate of individuals in courtrooms, AI’s integration promises a revolution in how justice is administered. In India, a country with a vast population and a complex legal system, AI’s integration is seen as a means to overcome deep-rooted inefficiencies, backlogs, and delays.
However, this potential also brings forth a host of ethical and legal questions about privacy, transparency, and human rights. This paper explores the promise and perils of AI to the various organs of the Criminal Justice System, particularly focusing on its application in India. PROMISE OF AI AI's capacity to process vast datasets, analyze patterns, and predict outcomes has led to its increasing use in various stages of the Criminal Justice process. Let us see the profound impact of AI on the Policing, Investigation, Prosecution, and Judiciary. a. Police Predictive Policing AI-Powered Predictive police tolls and techniques use demographics, historical crime information, and other contextual data to predict crime trends and their expected locations and times.
Example: The Delhi Police has piloted predictive policing programs, utilizing machine learning algorithms to anticipate criminal activity. Digital Surveillance AI-enabled surveillance systems can monitor large public spaces, analyze behavior patterns, and detect suspicious activities in real-time. The deployment of AI-based CCTV cameras in Hyderabad is a significant example of AI's role in surveillance and crime control. Traffic Control Automated traffic enforcement systems can detect and record traffic violations without human intervention. These systems can help reduce accidents and improve road safety. Sensors and Devices These AI-enabled devices can detect and identify illegal substances and their sources and can be used at border checkpoints, airports, seaports, and other high-risk locations.
Bots AI-powered bots interact with users to monitor conversations, detect suspicious patterns, and generate leads for law enforcement agencies. Fraud and Cybercrimes Detection AI can be leveraged to identify patterns and anomalies in financial and business data. AI can detect and block malicious activities such as hacking, phishing attacks, and malware. Speech Analysis AI-enabled text and speech analysis systems can monitor online and offline sources for hate speech and hate crimes. b. Investigation Forensic Analysis AI has made significant strides in forensic science, from facial recognition and fingerprint matching to DNA analysis.
Suspect Identification AI can analyze multiple data sources to determine the location and identity of suspects. AI can predict the best time and method to approach individuals. Interrogations AI can analyze verbal and non-verbal cues to detect deception or inconsistency. AI can improve the efficiency and accuracy of the interrogation process. Crime-Scene Reconstruction AI can integrate and synthesize evidence such as photos, videos, audio, and witness statements. AI can create a 3D or virtual reality model of the crime scene. Create Sketches AI can take the description of a suspect from a witness to produce a realistic image. c. Prosecution AI is transforming the legal profession in India, offering lawyers tools to improve efficiency, accuracy, and service delivery. AI allows legal professionals to focus on high-value strategic work while reducing costs and errors.
Faster Legal Research AI-powered research tools allow quick search through vast databases of case law, statutes, and legal precedents and help them save time and precisely identify relevant case laws. Automated Document Review AI tools can assist in reviewing legal documents such as contracts, petitions, or agreements and thus reducing errors and improves speed and efficiency. Predictive Analytics for Case Outcomes AI tools can predict case outcomes by analysing past cases and judgments thus allowing the councils to have a strategic plan and risk assessment. Legal Analytics and Insights AI tools can provide valuable insights into court decisions, judges' tendencies, and opposing counsel strategies. d.
Judiciary AI can support judicial decision-making by processing vast amounts of case law, legal precedents, and statutes to provide judges and lawyers with insights that can guide fairer and more informed decisions. For instance, COMPAS1 in the USA is widely used for such risk assessments, although it has raised concerns about fairness and transparency. In March 2023, HC of Punjab and Haryana used ChatGPT in a Bail order2. In the case of Desai3, the Supreme Court permitted the evidence recording through Audio-Video conference enlarging the scope of S. 2734 of Cr. P. C. to include the virtual presence. In the case of Anil Kumar Ghosh5, the Supreme Court stressed the importance of fairness in sentencing. AI could potentially help ensure more uniform decisions in line with principles of proportionality.
PERILS OF AI While AI offers numerous promises, its application in the Criminal Justice System is fraught with ethical, legal, and social concerns. Some of the major risks include bias, lack of transparency, accountability, and infringement on individual rights.
1 Correctional Offender Management Profiling for Alternative Sanctions COMPAS (software) - Wikipedia
2 Punjab and Haryana High Court uses ChatGPT in bail order (barandbench.com)
3 State of Maharashtra v. Praful B. Desai, AIR 2003 SC 2053
4 Evidence to be taken in presence of accused
5 Anil Kumar Ghosh v. State of W.B., AIR 2021 SC 2590
Deepfakes and Manipulation of Evidence Deepfakes are AI-generated media replicates real people, making it appear as though someone is saying or doing something they never did. Bollywood stars like Rashmika Mandhana6, Aamir Khan7 are victims of deepfake technology. Deepfakes raise several concerns: Ø False Evidence Creation: Deepfakes can be used to fabricate video or audio evidence, which could be introduced into trials, leading to wrongful convictions or false accusations. Ø Undermining Genuine Evidence: The existence of deepfakes could cast doubt on legitimate recordings, making it easier for criminals to claim that genuine evidence has been manipulated. Ø Legal System Overload: Courts will face the additional burden of evaluating the authenticity of digital evidence, which could slow down legal proceedings and increase litigation costs.
AI-Generated False Information AI technologies can be used to spread misinformation that may affect criminal investigations or legal proceedings: Ø AI-Generated Fake News: AI can generate highly believable fake news articles or social media posts that can mislead jurors, influence public opinion, or interfere with ongoing cases. Ø AI-Driven Social-Media Manipulation: In high-profile cases, AI can be used to flood social media with fake accounts or bot-driven narratives that attempt to sway 6 https://www. indiatoday. in/india/story/man-accused-in-actor-rashmika-mandannas-deepfake-video-casearrested-by-delhi-police-in-andhra-pradesh-2491281-2024-01-20
7 https://www.bbc.com/news/world-asia-india-68918330
public opinion or intimidate witnesses. Victim and Witness Harassment Ø Targeting Individuals: Deepfake technology can be used to create compromising or defamatory content targeting victims, witnesses, and legal professionals. This can be particularly harmful in cases of revenge porn or political retribution. Ø Intimidation and Harassment: Criminals may use deepfakes to blackmail or intimidate victims by threatening to release falsified media, further victimizing those already affected by a crime. Challenges with Digital Forensics As AI-generated content becomes more prevalent, the role of digital forensics in the criminal justice system will face new challenges: Ø Difficulty in Proving Authenticity: The proliferation of deepfakes, AI-generated texts, and altered digital content makes it harder for forensic experts to verify the authenticity of evidence.
Ø AI-Aided Counter-Forensics: Criminals may use AI to destroy or manipulate digital evidence, making it difficult for investigators to trace criminal activity. Algorithmic Bias and Discrimination AI systems, particularly those used in predictive policing and sentencing, often rely on historical data. If this data itself tainted with socioeconomic, racial, or gender biases, the AI system may reinforce and even amplify these biases. In the case of N. M. Thomas8, the hon’ble Apex Court examined the issue of systemic bias, highlighting that discrimination, even unintentional, needs to be addressed.
8 State of Kerala v. N.M. Thomas, AIR 1976 SC 490
Lack of Transparency and Accountability AI systems often operate as “black boxes,” where humans do not easily understand their decision-making. This lack of transparency is problematic in the Criminal Justice System, where fairness and accountability are paramount. Privacy Concerns and Surveillance Overreach Mass surveillance through AI-enabled systems can lead to a “surveillance state,” where individuals are constantly monitored, and their data is collected without consent. The landmark judgment in Puttaswamy9, Right to Privacy recognised as a basic fundamental right. The increasing use of AI in surveillance could infringe upon this right. Overreliance on Technology Overreliance on AI, without sufficient human oversight, could lead to erroneous outcomes, particularly when the data fed into AI systems is incomplete or flawed.
For instance, errors in facial recognition technology have led to false arrests in several countries. REGULATORY FRAMEWORK – AI SUPPORT IN THE NEW CRIMINAL LAWS IN INDIA As India moves forward with the introduction of new criminal laws – Ø BNS – Bharatiya Nyaya Sanhita Ø BNSS – Bharatiya Nagarik Suraksha Sanhita Ø BSA - Bharatiya Sakshya Adhiniyam It is critical to examine how AI can interface with these laws to improve efficiency, transparency, and effectiveness while ensuring that human rights and ethical standards are
9 Justice K.S. Puttaswamy (Retd.) v. Union of India, AIR 2017 SC 4161
upheld. Bharatiya Nyaya Sanhita The BNS replaced the IPC, introduced several reforms aimed at decolonizing the criminal justice system and addressing emerging challenges such as cybercrimes and terrorism. AI in Legal Research AI can assist in legal research, helping judges and lawyers sift through vast amounts of legal data to find relevant information. This can speed up trials and make the legal process more efficient, which is a central goal of the BNS and BNSS. AI in Cybercrime Regulation AI-driven cybersecurity systems can help prevent hacking, phishing, and data breaches, aligning with the BNS’s focus on combating new-age crimes like hate speech, radicalization, which are emerging threats to national security. Relevant Clauses supporting the Usage of AI Ø 78 – Stalking10 – added support for e-communication, Ø 111 – Organized Crime11 – added economic offence, cyber-crimes.
Ø 152 – Sedition12 – brought social media posts under this. Ø 196/197 – Hate Speech13 – added e-communication usage. Ø 294 – Obscene material14 – added revenge porn or violent videos. Ø 336 – Forgery15 – included e-records.
10 S. 78 of BNS, 2023
11 S. 111. BNS, 2023.
12 S. 152. BNS, 2023.
13 S. 196. BNS, 2023.
14 S. 294. BNS, 2023.
15 S. 336. BNS, 2023.
Ø 353 – Mischief16 –included e-records. Bharatiya Nagarik Suraksha Sanhita The BNSS replaces the CrPC. BNSS incorporates digital and forensic tools into law enforcement procedures, making room for AI-driven innovations to improve investigative accuracy and efficiency. AI in Policing and Crime Prevention AI can help law enforcement agencies in predictive policing, identifying crime hotspots, and deploying resources more efficiently. AI algorithms can predict where crimes are likely to occur, allowing the police to intervene proactively. In the famous case of Maneka Gandhi17, "just, fair, and reasonable" procedures established can apply to AI-driven policing tools. Predictive policing must align with principles of fairness, or it risks being challenged under this doctrine. AI in Investigations AI tools can analyze massive data, such as CCTV footage, much faster than humans.
AI can detect patterns in cybercrimes, frauds, and other complex cases where manual investigation would be time-consuming. AI algorithms can enhance images, transcribe conversations from poor-quality audio recordings, and even assist in DNA analysis. Relevant Clauses supporting the Usage of AI Ø 63 – Form of Summons18 – added support for e-summons. Ø 105 – Audio-Video Recording of Search and Seizure19 – introduced audio-video recording.
16 S. 353. BNS, 2023.
17 Maneka Gandhi v. Union of India, AIR 1978 SC 597
18 S. 63 of BNSS, 2023
19 S. 105 of BNSS, 2023
Ø 173 – FIR20 – information can be given thru e-mail. Ø 176 – Investigation21 – forensic evidence collection process shall be recorded on mobile phone or camera. Ø 183 – Recording of confessions and statements22 – shall be recorded through audiovideo on mobile. Ø 254 – Evidence for Prosecution23 – video conference support. Ø 530 – Trial and proceedings to be held in electronic mode24 – video conference support. Bharatiya Sakshya Adhiniyam The Bharatiya Sakshya Adhiniyam (BSA) is a reformed version of the Indian Evidence Act emphasizes the inclusion of digital evidence and recognizes the growing need for AI-driven tools in forensic analysis, data gathering, and evidence verification. AI in Evidence Management AI can streamline the process of collecting, storing, and analysing evidence.
Digital forensics can benefit from AI by automating data recovery from devices, ensuring chain-ofcustody protocols are followed, and verifying the authenticity of evidence. AI can help assess the credibility of witnesses by analysing their behavior and speech patterns, although this remains a contentious area. AI in Sentencing AI can also assist in sentencing by providing data-driven recommendations based on previous judgments and established legal standards. However, care must be taken to ensure that these recommendations do not perpetuate existing biases within the CJS.
20 S. 173 of BNSS, 2023
21 S. 176 of BNSS, 2023
22 S. 183 of BNSS, 2023
23 S. 254 of BNSS, 2023
24 S. 530 of BNSS, 2023
Relevant Clauses supporting the Usage of AI Ø 2(1)(d) – Document25 – elaborated to include all kinds of e-documents. Ø 2(1)(e) – Evidence26 – witnesses can now appear virtually before the court. The case of Desai27, allowed the use of video conferencing, highlighting the judiciary’s openness to technology. Ø 39 – Expert Opinion28 – widened scope of experts from “Any other filed”. Ø 57 – Primary Evidence29 – Recognizes electronic records as primary evidence. BALANCING THE PROMISE AND PERILS: THE WAY FORWARD The integration of AI into the Criminal Justice System is inevitable, but its implementation must be accompanied by robust safeguards to mitigate the risks while maximizing the benefits.
Six Principles for Responsible AI Corporations like Microsoft, Google, and Amazon, who are investing in building AI tools and technologies have proposed and following six key principles30 for responsible and trusted AI adoption: Ø Accountability Ø Fairness Ø Inclusiveness Ø Privacy and security
25 S. 2(1)(d) of BSA, 2023
26 S. 2(1)(e) of BSA, 2023
27 State of Maharashtra v. Praful B. Desai, AIR 2003 SC 2053
28 S. 39 of BSA, 2023
29 S. 57 of BSA, 2023
30 Responsible and trusted AI adoption - Cloud Adoption Framework | Microsoft Learn
Ø Reliability and safety Ø Transparency These essential basic principles are mandatory to create, use and adopt responsible and trustworthy AI tools and techniques in the mainstream products and services throughout the AI adoption journey. Establishing a Legal Framework for AI in Criminal Justice The Digital Personal Data Protection Act 202331 is a step in the right direction but not sufficient and must be complemented by laws that specifically address the integration and use of AI in surveillance, law enforcement, and judicial decision-making. Ensuring Algorithmic Fairness and Reducing Bias To mitigate algorithmic bias, AI systems must be carefully designed, tested, and monitored to ensure fairness. This requires diverse and unbiased training data, frequent audits of AI systems to identify and address biases.
Promoting Transparency and Accountability AI systems used in the Criminal Justice System must be transparent, with their decisionmaking processes open to scrutiny. AI algorithms used in Criminal Justice be interpretable and that their outputs are explainable. Accountability should be established to hold AI developers and users responsible for erroneous or biased outcomes. Safeguarding Privacy and Individual Rights The use of AI in surveillance must be balanced with the protection of individual privacy rights. Strict regulations should govern the use of facial recognition, video analytics, and other AI-driven surveillance technologies. Citizens should have the following rights:
31 Act No. 22 of 2023
Ø Right to Know32 how data is being used and have access to legal recourse in cases of privacy violations. Ø Right to Rectification33 of inaccurate data. Ø Right to Erasure (to be forgotten)34 of data. Ø Right to Object35 on processing of data.
CONCLUSION & SUGGESTIONS
AI has the potential to revolutionize the Criminal Justice System, offering solutions to long-standing challenges like inefficiency, backlog, and human error. However, the rapid integration of AI also brings with it significant ethical and legal concerns. To fully harness the promise of AI while avoiding its perils, careful regulation, transparency, and accountability are essential. As India continues to explore AI’s role in its Criminal Justice System, a balance must be struck between technological advancement and the preservation of fundamental rights and justice. The revolution of justice through AI is underway, but it is up to society to ensure that it is a revolution that enhances fairness, equality, and human dignity. Suggestions for Effective AI Adoption in the CJS
1) Establish Regulatory Framework for AI in Legal Practice
2) Mitigate Algorithmic Bias
3) Invest in AI Training and Legal Education
4) Develop AI Tools Tailored to Indian Law
5) Promote Access to AI-Driven Legal Services for All
32 Article 12, 13 and 14 of GDPR
33 Article 16 of GDPR
34 Article 17 of GDPR
35 Article 21 of GDPR
6) Ensure Privacy and Data Protection
7) Encourage Judicial Adoption with Safeguards
8) Create Accountability Mechanisms for AI Failures
9) Advanced Forensic Techniques
10) Legislation on Deepfakes