Artificial intelligence is transforming digital investigations. Tasks that once required hours of manual examination can increasingly be accelerated by systems that search, correlate and summarise large volumes of digital evidence.
But as organisations adopt AI-powered investigative tools, one question deserves just as much attention: are we developing the capabilities of our investigators as quickly as the technology they use?
The promise of investigative AI
Cellebrite’s Genesis illustrates the growing interest in purpose-built investigative AI. In his article “Cellebrite Genesis vs. ChatGPT: Why You Need Purpose-Built Investigative AI”, Justin Tolman discusses the importance of connecting AI-generated responses to source evidence, alongside the legal and operational requirements of investigations.
AI should help investigators reach the evidence faster — not relieve them of the responsibility to verify it.
— Cyber Squad Labs perspective on investigative AISource traceability matters because an investigative answer must be more than plausible. It must be testable against the underlying artefacts and evaluated in context. Even sophisticated tools cannot eliminate the need for investigative judgement.
The danger of confusing automation with expertise
Consider an AI system that identifies cryptocurrency transactions associated with a suspect’s mobile device. It may recognise wallet addresses, connect transactions and highlight relationships between entities. But does the investigator understand the difference between a wallet address and the person controlling it?
Can they distinguish a custodial exchange deposit address from a self-hosted wallet? Do they know where blockchain attribution is uncertain, and what additional evidence might be needed?
An AI-generated connection is not automatically proof of ownership, identity or criminal involvement. The same principle applies to mobile forensics, OSINT, communications analysis and other investigative disciplines.
Actual intelligence comes before artificial intelligence
Developing actual intelligence means giving investigators the practical knowledge to understand how evidence is created, recovered, interpreted and challenged. It requires more than learning which buttons to press.
Investigators need experience with unfamiliar data, competing hypotheses, misleading connections and situations where the evidence does not support the initial assumption. These skills become even more important when AI makes complex analysis appear effortless.
The missing investment: practical investigative training
Technology procurement often receives significant attention. Training can become a secondary exercise focused on platform features rather than investigative competence. Yet product familiarity and investigative competence are not the same thing.
A realistic multidisciplinary exercise might ask participants to:
- Examine digital artefacts recovered from a device.
- Identify and validate information through OSINT.
- Trace cryptocurrency transactions and test attribution claims.
- Correlate evidence from multiple independent sources.
- Develop and challenge competing investigative hypotheses.
- Explain their conclusions, methods and limitations.
AI should amplify expertise, not replace it
This is not an argument against artificial intelligence. Purpose-built investigative AI has the potential to reduce backlogs, surface overlooked connections and help investigators focus on decisions that matter. But speed and accuracy are not interchangeable.
Organisations should develop AI adoption strategies alongside equally ambitious plans for evidence literacy, investigative training and professional development.
Building the next generation of investigators
At Cyber Squad Labs, we believe effective training should reflect the complexity of real investigations. Digital evidence rarely exists in isolation: physical devices, cloud accounts, online identities, communications records and blockchain transactions can all contribute to understanding what happened.
Our approach connects these disciplines through realistic, hands-on investigative challenges that encourage critical thinking, evidence validation and independent problem-solving.
Before investing in artificial intelligence, invest in the actual intelligence of the people who will use it.
— Cyber Squad LabsFurther reading: Justin Tolman, Cellebrite Genesis vs. ChatGPT: Why You Need Purpose-Built Investigative AI, Cellebrite. Read the original article ↗. This is an independent Cyber Squad Labs commentary, not affiliated with or endorsed by Cellebrite. The article paraphrases the cited source rather than reproducing its wording.