Digital evidence is a cornerstone of modern federal investigations, used to adjudicate nearly 90 percent of all crimes committed today. The downside is that the expanding array of electronic data sources—from body-worn cameras and security devices to texts, emails, and social media—creates a massive amount of information.
For investigators, this presents a steep challenge: how to quickly analyze and prioritize case-related information to reach timely, accurate conclusions. Artificial intelligence (AI) can identify hidden connections between data points and spot minute details that could be missed by human analysts.
AI tools are designed to streamline digital workflows by automating essential but time-intensive tasks. Applying AI to investigative workflows within a framework of ethical and regulatory guidelines can be a force multiplier, enabling human decision-makers to act faster with more complete information.
How does AI empower investigators?
Digital evidence is a mix of words, images, sounds, locations, and numbers. Every day, more devices and sources add to an already growing mountain of data, including:
- Cell phone video and photos
- Social posts and electronic communication
- IoT devices, including smart TVs and appliances
- Automotive security platforms
- Low-cost video security systems
- New electronic methods of evidence capture
Emerging devices, such as AR (augmented reality) glasses, will only cause this information glut to balloon further.
Skilled analysts can find pertinent details buried in stacks of information, but the influx of evidence is fast outpacing the abilities of resource-constrained departments. AI-powered automation can sift through, organize, and uncover relevant information in a fraction of the time, enabling analysts and investigators to focus on the meaning behind the information.
For example, roughly half of law enforcement agencies in the US deploy body-worn cameras (BWCs), with 80 percent of large police departments using the technology. While most of this footage is tagged and stored in case it is pertinent to an investigation, AI solutions can help pinpoint relevant details when needed. Automation can translate, transcribe, and extract insights from BWC recordings in a matter of minutes—and can catch details the human eye may miss.
Reducing investigative time, expanding reach
Leveraging AI to augment investigative functions—from evidence management to language translation—enables investigators to react to intelligence in near-real time. In time-sensitive situations where intelligence grows less valuable with the passage of time, AI amplifies existing resources to accelerate investigations dramatically.
AI also enables accelerated data analytics at the edge. With AI capabilities now being embedded in smart devices and applications, tools like language translation and digital image enhancements are available on location—as are powerful analytics that can quickly make sense of disparate pieces of information.
The ability to conduct more thorough investigations while on location saves considerable time, which can make a difference in high-stakes situations. Live connectivity back to on-premises or cloud systems enables near-real-time data sharing for situational awareness. This, in turn, empowers field agents, investigators, and leadership to make faster decisions based on better information.
Trust in AI starts with trustworthy AI
Federal law enforcement agencies are actively looking to integrate AI into daily operational functions. Doing so responsibly starts with governance that ensures security, privacy, and ethical use of the technology. Trustworthy AI is more than ensuring the accuracy of the model and the system’s output; it is the result of guidelines and guardrails that protect the integrity of the information and how it is used.
Humans must validate AI-driven outputs before taking the next steps with that information. Agencies must also ensure that any AI systems they deploy have measures in place to eliminate bias, ensure data security, and restrict sensitive use cases that might be better suited to a human investigator.
In November 2023, Microsoft joined the Cybersecurity and Infrastructure Security Agency (CISA) and the UK National Cyber Security Centre (NCSC) along with 22 other domestic and international cybersecurity organizations in unveiling joint Guidelines for Secure AI System Development. These guidelines aim to advance the use of secure-by-design principles by providing stakeholders with recommendations on the secure design, development, deployment, and operation of machine learning and AI systems.
The transformational potential of AI dramatically improves the ability to conduct investigations and protect the public. Microsoft is committed to working with law enforcement agencies to refine not only the technology but also the guidelines and policies that ensure the responsible use of AI.
Discover more solutions that streamline and accelerate investigations and evidence analysis here.
