Thank you for the question. This is something we are actively exploring, but cautiously. My response reflects limited direct exposure to investigations and is largely informed by the Indian context.
Yes, a significant amount of time is spent on manual work such as reviewing large volumes of footage, identifying legally relevant moments, transcription, translation, redaction, and assembling evidence for lawyers, media, or campaigns. These are all areas where AI could plausibly reduce friction in the evidence-to-impact pipeline.
That said, a few constraints shape our thinking:
Risk and reliability matter more than speed—Investigations carry legal and safety risks, so any AI-assisted analysis needs to be highly reliable, explainable, and secure.
Context and judgment remain central—Deciding what evidence matters, how it fits into a pattern of abuse, and when it is strategically useful is still highly contextual and human-led.
Other bottlenecks are currently more binding—For many investigators, gaps in safety, mental wellbeing, legal support, training, and coordination matter more than the absence of advanced analysis tools, and premature AI use could even increase risk.
Where we see near-term promise is in assistive and investigator-controlled tools (e.g. secure transcription, translation, basic indexing, and first-pass flagging) that reduce cognitive load without replacing judgment.
Longer term, AI may help standardise evidence preparation so investigations are more consistently advocacy- and litigation-ready. For now, our priority is ensuring the surrounding infrastructure is strong enough that any future technical gains actually translate into impact.
We expect our thinking here to evolve as both the tools and the investigative ecosystem mature.
Thank you for the question. This is something we are actively exploring, but cautiously. My response reflects limited direct exposure to investigations and is largely informed by the Indian context.
Yes, a significant amount of time is spent on manual work such as reviewing large volumes of footage, identifying legally relevant moments, transcription, translation, redaction, and assembling evidence for lawyers, media, or campaigns. These are all areas where AI could plausibly reduce friction in the evidence-to-impact pipeline.
That said, a few constraints shape our thinking:
Risk and reliability matter more than speed—Investigations carry legal and safety risks, so any AI-assisted analysis needs to be highly reliable, explainable, and secure.
Context and judgment remain central—Deciding what evidence matters, how it fits into a pattern of abuse, and when it is strategically useful is still highly contextual and human-led.
Other bottlenecks are currently more binding—For many investigators, gaps in safety, mental wellbeing, legal support, training, and coordination matter more than the absence of advanced analysis tools, and premature AI use could even increase risk.
Where we see near-term promise is in assistive and investigator-controlled tools (e.g. secure transcription, translation, basic indexing, and first-pass flagging) that reduce cognitive load without replacing judgment.
Longer term, AI may help standardise evidence preparation so investigations are more consistently advocacy- and litigation-ready. For now, our priority is ensuring the surrounding infrastructure is strong enough that any future technical gains actually translate into impact.
We expect our thinking here to evolve as both the tools and the investigative ecosystem mature.