The Impact Of Artificial Intelligence On Forensic Science

INTRODUCTION

In our fast-paced, ever-evolving world, Artificial Intelligence (AI) is quickly becoming a game-changer across a wide range of fields, gaining more and more importance and recognition. Its impact is felt everywhere, from healthcare and transportation to finance and entertainment.

At its heart, AI is all about giving machines a touch of human-like intelligence, enabling them to understand, learn, and make decisions on their own in various situations. This means reimagining how we interact with machines, data, and the world around us. Decades of hard work and research have led to significant advancements in AI technologies. Think of things like machine learning, neural networks, natural language processing, and computer vision—they're all part of the AI toolkit. These tools help us analyse big datasets, spot trends, understand language, and make smart decisions, often without much human input.

Now, talking about forensic science—basically, using science to help out with legal stuff. The term "forensic" originates from the Latin word "forensis" which means "forum or a place of assembly. And its roots go way back to ancient Rome, where legal matters were settled in public forums. People would argue their cases, and decisions were made based on who had the best arguments and delivery.

In modern times, forensic science involves gathering physical evidence from crime scenes, analysing data, and reconstructing events to help bring justice. By carefully examining evidence, forensic scientists aim to uncover the truth and ensure fair legal outcomes.

Now, it is known that AI has permeated all the sectors, including the dynamic sector of forensic sciences, which is ever expanding. The intertwining of these two sects is studied meticulously and forms the basis of this article. 

ARTIFICIAL INTELLIGENCE AND IN FORENSIC SCIENCES

History

With the introduction of AI, forensic science has undergone a major transformation. 

The first recorded incident of AI in forensics dates back to the late 1980s. There was a groundbreaking moment when AI made its debut in 1989, in forensic science when the Federal Bureau of Investigation (FBI) introduced an AI system called the Integrated Ballistics Identification System (IBIS) to assist in the analysis of ballistic evidence. IBIS utilized image processing algorithms and pattern recognition techniques to compare images of bullet casings and identify potential matches, aiding in the investigation of firearms-related crimes. This groundbreaking use of AI was a big leap forward for forensic science. It showed just how much technology could do to make analysing evidence in criminal cases faster and more precise. It was like a shining example of how AI could really step up and help out in solving crimes more effectively.

However, the first incident of usage of AI in forensic sciences in India isn’t widely reported. And it is presumed that instead of a single ‘first incident’, there was a series of developments in this field which was gradual.

AI in Forensic Sciences

AI algorithms can process vast amounts of digital evidence—like emails, social media posts, and surveillance footage—much faster than humans ever could. They're also really good at spotting subtle patterns and anomalies in data, which can provide valuable insights that might be missed by human eyes.

AI also excels in tasks like facial recognition, voice analysis, and handwriting recognition, enhancing traditional forensic techniques. By leveraging AI's analytical abilities, forensic scientists can speed up investigations, improve the accuracy of evidence analysis, and ultimately help make justice more efficient and fairer for everyone involved. 

So, let us examine the various tools of AI in forensics-

a)    Facial Recognition and Biometrics

Facial recognition is a major tool in video forensic sciences in which a face is identified by analysing the facial features and comparing it with the virtual imaging against a database of photos to find the appropriate match. It uses various systems to compare facial features such as the distance between the eyes, nose, and mouth to a database of known faces to identify or authenticate individuals.

Every individual has a unique set of features that are also permanent, universal and inimitable in nature, such as fingerprint, iris, DNA, etc. All such features are collectively called biometrics of a person. Due to its uniqueness, it helps in identification and authentication of a person. Some are-

• Fingerprint Analysis- Automated fingerprint identification systems (AFIS) enables swift comparison of fingerprints against a large database of known individuals.
• DNA analysis: DNA can be obtained from a wide range of sources, including blood, semen, hair, and skin cells. It can also be used to exclude suspects from investigations.
• Voice recognition: It can be used to identify suspects by comparing the voice of a suspect to voice recordings found at a crime scene or on wiretapped phone calls.

Usage in Law Enforcement-

• Identifying Suspects
• Linking Suspects to Crime Scenes
• Exclude Suspects from Investigations

Evidentiary Value

In Indian law, biometrics generally fall under the category of primary evidence. 

Primary evidence refers to the original document or material object that directly proves a fact in issue. Secondary evidence, on the other hand, refers to evidence that is used to prove the contents of a document or the existence of a fact when the original primary evidence is not available or cannot be produced in court. It may include copies of documents, testimony of witnesses who have knowledge of the original document, or other forms of evidence that indirectly establish the fact in question.

In the case of biometrics, such as fingerprints or iris scans, they are considered primary evidence because they directly establish the identity of an individual, authenticate transactions, or support legal claims. Courts may rely on biometric data as direct proof of identity, provided that it meets the requisite standards of authenticity, reliability, and admissibility under Indian law.

Admissibility as per Indian Evidence Act, 1872- 

Biometric evidence can be admitted in court proceedings, provided it meets the criteria of relevance, authenticity, and reliability. The evidence should be directly linked to the facts in question and should be collected and preserved in an authentic and honest manner. 

Section 45 of the Evidence Act deals with opinions of experts regarding fingerprints. It states that courts may rely on the opinion of experts in determining the identity of individuals based on fingerprints. Expert testimony also strengthens the evidentiary value. Here it is interesting to note that, the qualifications and of the expert witness play a crucial role in determining the weight given to biometric evidence.

Also, when biometric evidence, like fingerprints or iris scans are backed up, with other types of evidence such as witness, or circumstantial evidence, it becomes more convincing. Having this extra backing makes the biometric evidence more reliable and compelling when presented in court.

Relevant Judicial Precedents

Bazari Hajam v. King Emperor (1921)- 

In this case, there was a question about the reliability of relying solely on fingerprints without any other corroborating evidence, to determine guilt. Bucknill, J., expressed concern about convicting someone of a serious crime based solely on fingerprint evidence without any supporting evidence. He also highlighted the potential for bias or unfairness in obtaining fingerprints from the accused for evidence. 

Fakir Mohamed Ramzan Vs Emperor (1935)-

In this case, the accused, was convicted on two charges under Sections 454 and 380 of the Indian Penal Code. The charges were related to two separate thefts. The peculiarity of the case is that there is absolutely no evidence to connect the accused with either of the offences except the evidence of the fingerprints. However, the court opined that although corroborative evidence is desirable, it cannot hold an accused guilty merely because of an opinion of the expert.  The court must satisfy itself as to the value of expert opinion in the same way as it must satisfy itself of the value of other evidence.

State of Maharashtra Vs Praful B. Desai (2003)- 

This addressed the admissibility of DNA evidence in Indian courts. The Supreme Court held that DNA evidence is admissible provided it meets certain pre-determined criteria that include proper collection, preservation, and analysis according to established scientific protocols.

State (NCT of Delhi) v Navjot Sandhu (2005)-

In this case, it was stated that secondary digital evidence could be adduced under other Sections of the Evidence Act, mainly Sections 63 and 65. This could be done irrespective of compliance with the requirements of section 65B and even without a certificate as specified in subsection 4 of 65B.

Anvar PV v PK Basheer and Others (2014)-

The decision in Sandhu case was reversed, and it was held conclusively that written and signed 65-B(4) certificate was mandatory for the admissibility of electronic records and no oral evidence could be adduced in support thereof. Thus, in the absence of such a certificate, it was argued that the VCDs could not have been admitted in evidence. It was confirmed in Arjun Panditrao that Anvar was the correct legal statement, and it could be concluded that section 65B was indeed the complete code on this matter.

Ritesh Sinha v. State of U.P (2013)-

This case explored the legality of voice exemplars. It was observed that hat the voice prints were being commonly used in other jurisdictions to detect crime. However, the court opined that the law, as it stood then, did not authorise the use of force to obtain a voice sample.

Tomaso Bruno Vs. State of Uttar Pradesh, (2015)-

In this case, it was established that CCTV footage is a strong piece of evidence. CCTV Footage is also admissible in evidence under Sec. 3 and 65B of the Evidence Act. For admitting this type of evidence also the requirements under Section 65B are to be satisfied. 

b)    Cyber Forensics and AI

Cyber Forensics and its Role in India

Cyber forensics, also known as digital forensics, involves the collection, preservation and analysis of digital evidence to investigate cybercrimes. It encompasses various aspects related to mobile devices, computers, and other electronic media. 

Its primary goal is to collect evidence that is to be presented in the court. For that, it goes through large amount of data and then preserves, identifies, extracts and documents the relevant data that is presented as evidence. 

Evidentiary Value

The Indian Evidence Act, 1872, was amended to include electronic evidence as part of the definition of “Evidence.”

Section 65A of the Indian Evidence Act is a key provision concerning electronic evidence. It expressly states that the contents of electronic records are deemed admissible in court proceedings.

Section 65B outlines the requisite conditions for the acknowledgment of electronic records as evidence. And, Sections 85A and 85B establish presumptions relating to certain electronic records and digital signatures.

Furthermore, Section 22A of the Indian Evidence Act was introduced to address the admissibility of electronic records as evidence in court. This section explains that data generated by a computer, stored, recorded, or reproduced on optical or magnetic media, printed on paper, or contained in an electronic record shall all be construed as documents within the legal framework.

Tools of AI in Cyber Forensics

With increasing usage of AI, it has increasingly been integrated into cyber forensic processes to increase accuracy. Here's how AI is intertwined with cyber forensics:

• Automated Data Analysis: AI can go through vast amounts of data more quickly and efficiently than human investigators. This is particularly important in cyber forensics, where large volumes of data are analysed from various sources to identify evidence.
• Pattern Detection: AI-powered tools can recognize patterns and anomalies in digital data that may indicate suspicious activities. For example, AI algorithms can detect unusual network traffic patterns or identify malware signatures in code, helping investigators pinpoint potential security breaches or cyberattacks.
• Natural Language Processing: NLP algorithms enable the analysis of unstructured textual data, such as emails, chat logs, and social media posts. It can then help investigators to identify key actors, motives, and communication patterns in cybercrimes.
• Image and Video Analysis: AI techniques enable the automated analysis of digital images and videos for forensic purposes. It can identify objects, faces, and activities in multimedia files which helps investigators to reconstruct digital evidence and understand the sequence of events in crimes.

Notable Case: 2008 Mumbai Terror Attacks

In the recent past, digital forensics played a crucial role, during the 2008 Mumbai terror attacks. It was used in analysing communication networks and electronic data.

Investigators examined phones, emails, GPS data, and internet usage. Such a meticulous extraction and analysis of information helped to reconstruct the perpetrators’ activities, communication channels, and coordination patterns. Cyber forensic experts were able to piece together a timeline of the attackers' movements. They scrutinized digital footprints to identify the perpetrators' networks, uncovering crucial insights into their planning and execution of the attacks.

Cyber Forensic Labs in India

Recognising its importance India has taken steps to enhance its cyber forensic capabilities. 18 cyber forensic-cum-training laboratories have been commissioned across the country to combat cybercrimes, including financial frauds and crimes related to women.

c)    Predictive Policing and AI

Predictive policing is said to be a positive strategy that involves forecasting the various aspects of earlier crimes such that it can be helped to prevent potential future crimes.

The dual purpose of predictive policing is-

• Crime Prevention
• Resource Optimization

The traditional method of policing is reactive in nature. This means the law enforcement agents react to crimes after their incidence. However, with the use of AI we can now analyse the past records and data to detect a pattern and then eventually forecast where and when crimes are likely to occur with tools such as data analysis, statistical modelling, etc. This is therefore predictive or futuristic in nature.

Predictive Policing can be divided into 4 different categories-

• Those which predict crimes
• Those which predict offenders
• Those which predict offender’s identity
• Those which predict victims

Adoption and Cases in India-

In the early 2000s, Indian police forces started embracing innovative approaches like data mining and analysis to anticipate and prevent crimes. 

• JARVIS- It stands for Joint AI Research for Video Instances and Streams. It is a video analytics platform that was introduced by the start-up Staqu in November 2019. This software's goal is to produce valuable data from lengthy CCTV video footage with brief, clear real-time notifications utilising AI and computer vision, greatly cutting the time it takes to produce useful data. It thus assists law enforcement agencies in tracking each violent occurrence that occurs in a certain location and the police be able to mobilise officers to prevent any escalation of the incident with the help of such real-time event identification
• CMAPS, Delhi- Among the early adopters was the Delhi Police, known for its pioneering efforts. In 2015, they rolled out the Crime Mapping, Analytics, and Predictive System (CMAPS), a cutting-edge initiative that taps into real-time data from the city police’s helpline to pinpoint areas prone to criminal activity. It collects data every 3 minutes from the Indian Space Research Organization’s satellites. This helped in identification of crime hotspots.
• CCTV Matrix, Himachal Pradesh- On January 1, 2020, Himachal Pradesh police installed over 19,000 CCTV cameras throughout the state. They've named this extensive network the "CCTV Surveillance Matrix," which lays the groundwork for an innovative predictive policing strategy. Their vision is to expand this network to a remarkable 68,000 CCTV cameras, with the aim of having nearly one camera for every 100 residents.

There are looming concerns of effectiveness as well as privacy rights which are discussed further in this article.

CHALLENGES AND ETHICAL CONSIDERATIONS

Despite its recent advancements and widespread usage, there are certain loopholes in the forensic AI systems. Notwithstanding the transformative potential of AI in forensic science, its adoption poses multifaceted challenges and ethical dilemmas, some of which have been discussed below-

• Privacy and Data Risks- 

As per our study, we know, predictive policing algorithms use vast amount of personal data, I must raise the effect they have on the right to privacy. The opacity regarding the use of personal data is a violation of the right to privacy as envisioned in Justice K.S. Puttaswamy (Retd.) v. Union of India (2017). With increase in capability of AI, finding out and analysing minute details, the risk of disregarding individual’s privacy escalates.

• Consent Concerns-

In the realm of forensics and AI, there's a big concern about consent. Imagine if your personal data, like your fingerprints or images from surveillance cameras, were used by AI without your permission. This raises important questions about privacy and fairness. People should have a say in how their information is used, especially when it's used by AI to make important decisions. So, in forensics, which includes usage of mechanisms such as CCTV and compulsory biometric sharing. it's really important to make sure that people agree to how their data is being used, to make sure everyone is treated fairly and respectfully.

• Algorithmic Biases- 

Skewed training data or flawed algorithms, can cause algorithmic biases and can spread disparities in criminal justice outcomes, particularly affecting marginalized communities. Predictive policing largely relies on past historical data which indicates who is more policed instead of who is more likely to commit the crime. And as per records, it is the marginalised minorities who have been systematically targeted by the law enforcement agencies. So, usage of such system exacerbates the existing societal inequalities.

• Black Box Nature- 

AI is called to be opaque in nature. It crunches a bunch of data and gives results, but it doesn't always explain how it came to those conclusions. This lack of transparency worries people because it means we might not fully understand why AI makes certain judgments. And when it comes to things like criminal trials, where someone's freedom is on the line, transparency is super important. We need to know how AI reached its conclusions so that people can be held accountable, and everyone gets a fair shot at justice.

• Lack of Comprehensive Legislation- 

Right now, in India, there aren't definite rules just for AI and its usage in forensics. Certainly, there are some bits in other laws, like the Information Technology Act, 2000, and the Personal Data Protection Bill from 2019. But these laws don't actually cover everything about AI. They touch on some parts, but they don't really dive into all the tricky stuff that comes with AI. We need laws that really understand AI and can handle all the different things it can do, from robots to algorithms.

• Lack of comprehensive regulatory oversight-

The absence of a detailed regulatory authority for AI can result in uneven oversight and limited enforcement of AI-related regulations. The absence of well-defined and enforceable ethical guidelines for AI development and usage in India poses a challenge. This dearth of comprehensive guidelines may lead to inconsistent practices and potential misuse of AI systems.

Despite the transformative potential of AI in forensic science, its adoption poses complex challenges and ethical dilemmas. Legal frameworks must grapple with the admissibility and reliability of AI-generated evidence, ensuring adherence to principles of fairness, reliability, and procedural justice.

COMPARATIVE ANALYSIS

India

India has been gradually integrating AI into forensic sciences over the past decade, with significant advancements seen in recent years. Even then, it can be said that AI adoption in forensic sciences is still in its nascent stages in India compared to some other countries

AI is being used in India for tasks such as facial recognition, fingerprint analysis, voice recognition, and crime pattern analysis. It is also utilized for enhancing surveillance systems and managing large volumes of digital evidence.

Government agencies and research institutions are collaborating to develop more sophisticated AI solutions tailored to Indian forensic requirements.

Prominent Project- Automated Fingerprint Identification System is developed by the National Crime Records Bureau (NCRB), AFIS uses AI algorithms to match fingerprints from crime scenes with a national database that aids in criminal investigations.

United States

The US has been at the forefront of AI adoption in forensic sciences for several decades, with continuous advancements in technology and methodologies. It is deeply integrated into various aspects of forensic sciences in the US and plays a key role in crime investigations, evidence analysis, and courtroom proceedings.

It is used for facial recognition, fingerprint matching, handwriting analysis, DNA sequencing, and ballistics analysis. It is also used in cyber forensics for investigating digital crimes.

Recent developments include the use of machine learning algorithms for analysing complex forensic data and the deployment of AI-powered tools for real-time crime detection.

Prominent Project- Patternizr is developed by the New York Police Department (NYPD). It is an AI-powered tool that analyses crime patterns and connects seemingly unrelated crimes, assisting detectives in identifying serial offenders.

Japan

Japan has been exploring the use of AI in forensic sciences in recent years, with a focus on leveraging technology to improve investigative processes and enhance public safety. While AI adoption in forensic sciences is still emerging in Japan, there is a growing interest among law enforcement agencies and research institutions to explore its potential applications.

It is used in Japan for facial recognition, handwriting analysis, voice identification, and analysing digital evidence in cybercrime investigations. Additionally, it is also utilized in forensic pathology for medical examinations and cause of death determinations.

Japan is investing in AI research and development for forensic sciences, with efforts focused on developing advanced algorithms for evidence analysis, enhancing forensic databases, and improving collaboration between forensic experts and technology developers.

Prominent Project- The National Research Institute of Police Science (NRIPS) is developing AI algorithms for automated handwriting analysis. It enables quicker and more accurate examination of handwritten documents in criminal investigations.

United Kingdom

The UK has been gradually adopting AI in forensic sciences over the past decade, with increasing emphasis on innovation and collaboration between academia, industry, and law enforcement agencies. And it is becoming an integral part of forensic investigations in the UK, with growing recognition of its potential to enhance efficiency and accuracy in evidence analysis.

The UK is investing in research and development of AI technologies tailored to forensic applications, including advancements in biometric identification, speech recognition, and video analysis. Collaborative initiatives are underway to standardize AI usage in forensic practices and ensure compliance with legal and ethical guidelines.

Prominent Project- Project ECHO is developed by the Metropolitan Police Service. It utilizes AI algorithms to analyse digital evidence from multiple sources, including smartphones and computers, to expedite investigations into cybercrimes and online fraud.

FUTURE TRENDS AND DEVELOPMENTS 

As we look ahead to the future of AI in forensic sciences, it's like stepping into a world full of exciting opportunities. Picture this: crime scenes being investigated by smart machines, powered by quantum technology that can crack even the toughest codes. But as we venture into this realm of innovation, we mustn't forget to consider the human side of things—the ethics, laws, and how society will be affected by these new AI-driven forensic tools. Let us explore the various up and coming tools-

• Augmented Reality Visualization: 

Augmented reality (AR) technologies are expected to revolutionize the picturing of forensic evidence. Imagine forensic analysts donning AR headsets, overlaying virtual crime scene reconstructions onto real-world environments, and visualizing forensic data in three-dimensional space. This may enhance collaboration among forensic teams, facilitate evidence interpretation, and present compelling visualizations in courtroom proceedings.

• Quantum Computing for Cryptanalysis: 

Quantum computing in cryptanalysis, offers unmatched computational power to decipher encrypted data. In forensic sciences, quantum computing is capable of unlocking encrypted communications, decrypting digital evidence, and thwarting cybercriminal activities. 

• Blockchain for Chain of Custody: 

Blockchain technology is known for its immutable and transparent nature. This holds great promise for maintaining the chain of custody in forensic investigations. By recording the transfer of evidence custody in a secure and decentralized ledger, blockchain ensures the integrity and authenticity of forensic evidence throughout its lifecycle. 

To make sure we're on the right track, it's crucial that we approach AI deployment with responsibility. This means being mindful of the impact our technology can have and working together across different fields to address any concerns that arise. By doing so, we can unlock the true potential of AI in forensic sciences and ensure that justice continues to prevail, even in this digital age.

CONCLUSION

In conclusion, the fusion of artificial intelligence with forensic science marks a remarkable leap forward in innovation and effectiveness. From recognizing faces to investigating cybercrimes and predicting patterns in policing, AI-powered tools have revolutionized traditional forensic methods, making investigations faster, more precise, and data-focused. Yet, while AI brings immense benefits, it also presents various challenges concerning bias, privacy, lack of legislation and regulatory oversight and ethical dilemmas that require careful consideration.

As technology progresses, it's crucial to deploy AI responsibly, prioritize transparency, and uphold ethical standards. This ensures we harness its transformative power in forensic science while mitigating potential risks. Through ongoing research, collaboration, and the establishment of regulatory guidelines, AI holds the promise of fundamentally reshaping forensic science, bolstering justice and security in our increasingly digital world.

FREQUENTLY ASKED QUESTIONS

1.    Is evidence obtained through AI analysis admissible in court?

In many jurisdictions, evidence obtained through AI analysis is admissible in court, but its admissibility may be subject to judicial scrutiny. Courts typically assess the reliability, relevance, and fairness of AI-generated evidence before admitting it, considering factors such as the validity of the underlying algorithms, the quality of the data, and adherence to legal standards.

2.    How do courts evaluate the reliability of AI algorithms in forensic analysis?

Courts may evaluate the reliability of AI algorithms by considering factors such as the methodology used to develop the algorithms, the accuracy rates demonstrated through validation studies, the transparency of the decision-making process, and any known limitations or biases. Expert testimony and scientific evidence may be presented to assist the court in assessing the reliability of AI-based forensic methods. Further, availability of corroborative evidence strengthens the case.

3.    Are there legal standards for validating AI algorithms in forensic analysis?

No, there are no comprehensive set of legal standards for validating AI algorithm in forensic analysis. However, the unofficial legal standards may vary by jurisdiction, courts generally expect rigorous validation studies that show the accuracy, consistency, and generalizability of AI-based methods.

4.    Can defence attorneys challenge the admissibility of AI-generated evidence in court?

Yes, defence attorneys can challenge the admissibility of AI-generated evidence in court by raising objections based on factors such as the reliability of the AI algorithms, the quality of the data used, the lack of transparency in the analysis, or the potential for bias or error. Courts may conduct hearings or require expert testimony to address these challenges and determine the admissibility of the evidence.