Hours of struggle, stressed and tired mind, and fear of errors, all replaced by the Artificial Intelligence. Within a snap of time, with every minute and precise detail, the Artificial Intelligence smoothly generates the outcome.
Artificial Intelligence, these days are trusted and often used to treat complex problems and not only the complex ones but every other thing, which is time consuming. Every crucial field has a touch of Artificial Intelligence in it, whether it’s the field of medical, automobiles, science, computer science or forensics. Artificial Intelligence or in abbreviation called “AI”, imitates the human intelligence.
AI and forensic science
Artificial Intelligent not only imitates human brain but is likely more advanced and prudent when it comes to the field of law. Recently, there are many instances where AI was used by law enforcement agencies such that for surveillance at Ram Temple in Ayodhya. After which Ahmedabad became the first city to get AI linked surveillance systems. However, in the most acknowledged manner, AI was used to solve the blind murder case, where by restructuring the face of murder victim using the facial recognition process the perpetrators were apprehended.
Similarly, in the present times, where AI is overtaking all the possibilities, it becomes essential to evaluate the potential of AI in the field of forensic science and to analyse the positive and negative outcomes through which crime solving become faster and accurate.
Aiding of AI
- Digital Forensic :
Digital forensics is tremendously evolving with the help of AI. Because of which forensic scientists are at a lot of ease.
As the major objective of forensic scientists is to examine, analyse evidences, assist in the investigation and to prosecute the perpetrators of crime or to absolve an innocent person from suspicion. In accordance to these objectives, AI, analyse and simplify the data, which ease the load of these scientists, and assists the investigation with speedy and precise conclusions and acts as a helping hand in crime fighting.
- AI tools:
There are many tools which are AI powered and are used worldwide to reduce the crime rate in different areas. Every year, thousands of tons of sandalwood are smuggled from India, at such times when the wood is seized, an AI based identification device named “Xylotron” shall be used to quickly identify the wood rather than sending it to forensic laboratories. Some other devices like “A.A.R.I.” (Amino acid rapid imager), which contains AI based Ridge detection software tools, which eliminates the intensive labour and increases speed when it comes to finger print detection shall be brought in practise.
- Argumentation and graphical construction:
AI models are programs that detect specific patterns using a collection of data sets. It is an illustration of system that can receive data inputs and draw conclusions or conduct actions depending upon those conclusions. These AI models mainly work on argumentation and scenario construction methods, where the graphical representations
can be made by methods like NLP, by using statistical evidence which will ultimately help in drawing an accurate conclusion for speedy justice.
- Pattern Recognition:
Recognising patterns is one of the most crucial aspect and objective of forensic science. AI through its capacity to analyse Meta data by scanning through all the compound databases easily breaks and recognizes the patterns. Pattern recognition can be performed in image recognition or in text recognition such as hand writing recognition and even in sound recognition. The power of AI in the area of pattern recognition makes the process less time consuming and simplifies the complexities of comparing the subject matter with all the existing data sets.
- Common medium among all authorities:
Often, miscommunication takes place between Advocates, Forensic Investigators and the Judges, when it comes for the interpretation of any evidences in any case. Such incidents, hampers, marshalling and appreciation of evidences, affecting the ratio decidendi of the case and leads to grave injustice.
However, through AI these gaps can be filled, as AI provides a common medium among all the authorities which reduces the risk of any such errors.
- Storehouse and Data mining:
In modern times where it has become difficult to store large amounts of data as when the data storage exceeds the maximum capacity then forensic institutions faces problems like shutting down of systems, errored information and much more. So, to minimize such issues AI’s storage capacity can be utilised to store the large amount of data of different cases as well as the laboratory findings and the results and all other information can be stored.
By using such AI powered Data houses any forensic investigator just by adding a prompt can highlight or access any piece of information from all the files and
documents. This makes the process of data mining easier and faster in comparison to going through all the data manually.
Challenges
- False outcome:
Artificial Intelligence, purely works upon the basis of data fed to it, but at instances it creates ruckus by providing wrong outcomes. For example, the COMPAS algorithm for predicting recidivism has been found to reinforce problematic racial and social biases. Similarly, in 2018, it was found that facial verification and identification technologies, which are used by the law enforcement, generates high errors in a disproportionate ratio for the racial minorities. AI no doubt makes the identification process easy by limiting the options but often it creates errors and leads to wrong convictions, as it does not consider the physical features like height and weight while analysing. Hence, false outcomes are one of the greatest drawbacks of AI.
- Expensive process:
Artificial Intelligence, no doubt reduces the manual work but at the same time it is very costly. As in digital forensics with advancements, development arises, and to update and maintain the technology, more cost is needed[11], which becomes an issue. Many enforcements agencies struggle to have an adequate budget for implementing AI technology into their forensic procedures.
- Evidentiary value:
AI generated outcomes have a disputable evidentiary value, as they are less reliable and are nowhere expressly recognized by the legislature.
Admissibility of evidences is regulated under sec 45 to 51 of the Indian Evidence Act, 1872. Where sec. 45 states that when the court lack expertise in any matter than it shall frame its opinion on the basis of the “expert’s opinion, (a person who is skilled in the field)”. Whereas the language of sec. 51 of the Evidence Act too focuses upon the opinion of any living person. However, as per sec. 11 of the Indian Penal Code, 1860, AI does not qualify the definition of person. Whereas, as per sec 45-A, when the evidences are in electronic form, then an examiner who is an expert in electronic record shall be present in the court to testify the evidence.
AI works on limited information accessed by it, which may not always corroborate the facts precisely, as often it may not support any fact that strongly, or may remain silent upon any aspect of any fact, which may increase the chances of injustice, hence on evidentiary perspective, it is less reliable.
- Complex in updating data:
Analysing data is only possible after feeding the sufficient and adequate quantity of data, i.e. it should neither be overfitting nor under fitting in order to have a proper outcome for which fine tuning of algorithms and hyper parameters is done, which again is a complex process. For generating precise data pre-processing steps too needs to be done and along with-it dimensionality reduction techniques and algorithm selection needs to be done. All these processes, makes it complex to update the data in AI.
Conclusion
AI, in a very short span of time has become almost a part of every field, especially in forensics. No doubt, it is tremendously serving in all the fields but till yet, there is no any specific law to govern every aspect of it, which in near future, especially in areas where forensics and judiciary are interconnected and the matter of rights are involved,
the parliament must, keeping all its drawbacks in view, shall frame a specific legislature to regulate it.
Authors:
Author:
Hetkumar Joshi,
Student,
4th year, BBA LLB
Navrachana University.
Co-author:
Taiba Walia Makrani,
Student,
4th year, BBA LLB
Navrachana University.
[1] Rohit Kumar Singh, Future-ready AI surveillance likely to guard Ram temple in Ayodhya, HINDUSTAN TIMES, (Jan. 04, 2024), https://www.hindustantimes.com/cities/lucknow-news/futureready-ai-surveillancelikely-to-guard-ram-temple-in-ayodhya-101704312470074.html.
[2] MD Ijaj Khan, AI is watching you! Ahmedabad becomes the first city to get AI- Linked Surveillance system, HINDUSTAN TIMES TECH, (Jan.02, 2024),https://tech.hindustantimes.com/tech/news/ai-is-watching-youahmedabad-becomes-first-city-in-india-to-get-ai-linked-surveillance-system-71704199403087.html.
[3] How AI help Delhi Police to solve a Blind Murder case, THE ECONOMICS TIMES, (Jan. 25, 2024), https://economictimes.indiatimes.com/news/new-updates/how-ai-helped-delhi-police-to-solve-a-blind-murdercase/articleshow/107122601.cms?from=mdr.
[4] Alex C. Wiedenhoeft, Ph.D, John C Hermanson , The Xylotron: A Field-Deployable Machine-Vision Wood Identification System, FOREST SERVICE, US DEPARTMENT OF AGRICULTURE, (Mar. 21, 2022), https://www.fs.usda.gov/research/news/highlights/xylotron-field-deployable-machine-vision-woodidentification-system.
[5] FOSTER+FREEMAN, https://fosterfreeman.com/aari/. 6HEWLETTPACK,https://www.hpe.com/in/en/what-is/aimodels.html#:~:text=AI%20models%20or%20artificial%20intelligence,actions%20depending%20on%20those %20conclusions.
[6] ORACLE,https://www.oracle.com/in/artificial-intelligence/what-is-natural-languageprocessing/#:~:text=Natural%20language%20processing%20(NLP)%20is,natural%20language%20text%20or% 20voice.
[7] LEXIS NEXIS,https://www.lexisnexis.co.uk/legal/glossary/ratio-decidendi.
[8] EURONEWS.NEXT, https://www.euronews.com/next/2023/11/20/unreliable-research-assistant-false-outputsfrom-ai-chatbots-pose-risk-to-science-report-s.
[9] Sarah Barrington and Hany Faird, A comparative analysis of human and AI performance in forensic estimation of physical attributes, SCIENTIFIC REPORTS (Mar. 23, 2023), https://www.nature.com/articles/s41598-02331821-3.
[10] Supra note 11.
[11] Louisa S., How does AI Contribute to Digital Forensics, LINKEDIN (May. 16, 2023), https://www.linkedin.com/pulse/how-does-ai-contribute-digital-forensics-louisa-fornah-she-her-/.
[12] The Indian Evidence Act, 1872, § 45, No. 1, Acts of Parliament, 1872 (India).
[13] The Indian Evidence Act, 1872, § 51, No. 1, Acts of Parliament, 1872 (India).
[14] Indian Penal Code, 1860, § 11, No. 45, Acts of Parliament, 1860 (India).
[15] The Indian Evidence Act, 1872, § 45-A, No. 1, Acts of Parliament, 1872 (India).
[16] Navin Chanchal, Navigating the Complexities of AI Algorithms: Unlocking the Power of Artificial Intelligence, LINKEDIN(Sep.16,2023),https://www.linkedin.com/pulse/navigating-complexities-ai-algorithms-unlockingpower-chanchal/.