ARTICLE
7 October 2024
Getting AI to do the heavy lifting
Article by Dr Huw Bristow, Altia CTO. This article was originally published in SOCEX Strategic event brochure.
Sometimes when you think or hear of AI, everything gets a bit ‘sci-fi’ which, depending on your perspective, can be very exciting or quite alarming!
But, as we’ve talked about before at SOCEX, there’s a whole host of every-day things we interact with that are powered by AI. They are often quite mundane, like deciphering the washing labels on clothes, but probably make more ‘real’ difference than headline grabbing technology advances.
The use of AI in policing is proving to be much the same. At Altia, two of our most recent uses of AI, is generative AI and deep learning which is incredibly powerful when it comes to interpreting large data sets. It’s the difference between laborious manual data entry and analysis, and streamlined, efficient operations – releasing the time of precious human expertise to areas of greatest value.
To take a specific example, transcribing interviews and analysing the resulting data used to be a time-consuming task. But deep learning now enables in-house real-time transcription and can extract crucial information quickly and cost-effectively. AI also facilitates entity matching, swiftly identifying individuals, locations and other pertinent details in vast datasets, thereby supercharging case management. Another powerful capability of generative AI is summarisation. It takes large amounts of textual data and summarises it quickly which is particularly crucial in live crime situations.
Without a doubt, generative AI’s ability to find information (especially when it’s a needle in a haystack) while also providing links to the original source, is transformative. And, as well as speeding up manual processes or finding new information, generative AI also aids in transparency by capturing and searching sources and automating the process of recording decisions made. This all reduces the administrative overhead of tracking and reporting on audit trails.
Another increasingly prevalent method of AI is its transformation of interaction techniques. Natural language processing, also a component of generative AI, enables users to make inquiries of systems in a manner that feels intuitive and familiar. For instance, users can request, “List the top ten transactions in my bank statements.” This aligns with our natural preference for conversational interactions. We feel this has particular relevance when speed is of the essence in live crime instances, and with users who aren’t necessarily system super users. AI context and continued conversations has improved even further over the last 12 months and again improves interaction ease and familiarity.
In summary, the question we ask, at Altia, of every AI development we focus on is, ‘What does it do better that is already being done, and how can we make it trustworthy?’ This means the right controls are already in place, and it’s clear where AI can be safely applied to turbo charge the process with information and, or, pace. The journey of AI has just begun. I am confident it promises a safer, more secure and efficient future. It is set to be transformative for all industries – not just policing – as long as we have the right human guidance and control around it. We don’t want it getting all Sci-Fi on us after all!
