Verification Key as AI Enters Crime Investigation, Say Experts
Lucknow, October 23: Artificial intelligence (AI) could become a major asset in forensic science and criminal investigation, making the process faster, more effective and more accurate. Experts say the technology can be applied to fingerprint comparison, DNA data analysis, CCTV footage scrutiny, digital evidence testing and the identification of connections between persons and incidents linked to crimes.
Yet, they warned, relying on AI results without independent verification may expose investigations and judicial proceedings to serious risk.
Speaking at a panel discussion on 'Artificial Intelligence in Forensics and Investigation', specialists said AI is a useful supporting tool but not a replacement for expert judgement, human reasoning or independent validation of evidence. The discussion took place on Friday at the Uttar Pradesh State Institute of Forensic Science, as part of the 'TechSangam' session of the three-day international event Samvad-26.
The session opened with remarks from moderator Ashish Tiwari, DIG (Technical), who welcomed the participants. He spoke about how forensic science contributes to the prevention, detection, investigation and prosecution of crime. He also referred to the provisions of the Bharatiya Nyaya Sanhita (BNS) and the significance of forensic evidence within the wider justice system.
Summing up the discussion, panellists agreed that AI can become a potent technical instrument for investigators, provided its use rests on credible data, transparent methods and human oversight. They said that as investigations adopt faster tools, safeguarding the authenticity of evidence and the trustworthiness of the legal process remains equally vital.
Sandeep Joshi noted that AI can analyse large datasets within a short period, accelerating investigations that would otherwise take considerable manual effort. Patterns and possible associations in information can be recognised swiftly, giving investigators key details in time to make better decisions.
He listed several potential uses. AI can help compare fingerprint patterns, process intricate DNA data, detect suspicious activity in CCTV footage and trace the movement of people across different places. It can also examine large volumes of forensic data to uncover likely links between individuals, incidents, locations and evidence.
Joshi cautioned that AI effectiveness is shaped by data quality, suitable software and hardware, and the analytical methodology. For dependable outcomes, he said, attention must be paid to the process of using the technology as much as to the technology itself. The panel also discussed questions about the reliability of AI in detail.
Experts observed that while AI can strengthen investigations, incorrect or biased data, misleading conclusions and tampering with digital evidence may create significant challenges. AI can sometimes produce findings that are not adequately backed by available material. Biased training data can also skew results, and the spread of deepfakes and other artificially created content increases the threat of misleading evidence.
According to the speakers, AI findings must be independently confirmed, original evidence must be checked for authenticity and the entire analytical process must be recorded. Investigators should compare AI outputs with other reliable sources and traditional forensic methods. Human supervision, accountability and clear legal and ethical standards are essential in the use of such technology.
The discussion also highlighted the principle of 'AI for security and security of AI'. Experts said AI will benefit investigations only if systems are protected against errors, tampering, misuse and unauthorised access.
Sandeep K. Satyapathy (retired) said effective use of AI in forensic investigation requires deep subject knowledge, practical experience and technical understanding. Investigators must know how to frame the right questions, assess the quality of available data and identify possible errors or misleading information in AI outputs. He referred to his professional association with Dr. G.K. Goswami and to an episode from the 2006 Nithari case in Noida.
He also recalled a conversation with journalist Karan Singh and stressed the value of guidance, practical experience and accumulated expertise in investigative work. Satyapathy said AI can analyse information and suggest possible conclusions, but it cannot replace the professional judgement of an investigator.
To assess the reliability of AI outputs, he said, attention must be paid to how evidence is collected, the chain of custody, the methodology of analysis and the verification of results. Admissibility of AI-assisted findings in court depends on applicable legal provisions and the credibility of the evidence. An AI-generated conclusion does not automatically make evidence admissible.
The relevant expert must be able to explain the analytical method, confirm the results and establish their relevance and reliability. Responsible use of AI in forensic investigation and judicial proceedings, the speakers concluded, requires human oversight, professional expertise and proper validation of evidence.