The past

When I first learned how to do research, everything happened in person. You walked into the library, pulled one of those long wooden drawers from the card catalog, and flipped through hundreds of little cards until you spotted the right one. Those drawers seemed to stretch forever. It always amazed me how far they pulled out, as if they somehow defied the laws of physics.

Eventually, the shelves got replaced by digital catalogs and online databases. What once took hours now took minutes. Fast-forward, and we’ve entered another new phase, one where artificial intelligence (AI) doesn’t just store information but gathers and helps digest it for us.

I teach criminal justice research methods at a university, and my students are always shocked when I tell them I’m older than Google, which was created in the late 1990s.

The future

I suspect what will surprise future generations is that we used Google and other search engines by writing in keywords and getting thousands of websites, and we would tediously visit these various websites and try to find what we needed.

In light of AI, we no longer have to painstakingly visit a ton of different websites to find what we need because AI will search, locate and synthesize the information for you. What’s more, AI will even give its sources if you want to dig deeper or verify the information yourself.

In sum, finding information and the idea of understanding research have changed.

In light of advancements in AI, here are a few ways to make sure we’re finding, interpreting and applying that information in ways that serve our departments and communities.

Why evidence still matters

Consider the following hypothetical example. A local politician proposes bringing back a Scared Straight program for youth offenders. It sounds convincing to him. Show kids the consequences of prison life, and they’ll straighten up. But CrimeSolutions.gov, run by the Department of Justice, rates Scared Straight-type programs as ineffective and even harmful by leading to higher rates of recidivism.

What makes the site worth bookmarking is its process. Every program, whether it’s crime prevention, law enforcement technology or school violence, is reviewed by subject-matter experts. Each one lists the actual peer-reviewed studies that informed ratings of effective, promising, ineffective or negative effects so you can trace the reasoning to the research. For a department leader or training coordinator, that’s golden. It means you’re not just told what works. You can look at the research and see why.

When research gets it wrong

However, even peer-reviewed studies deserve a second look. Peer-reviewed studies are an important safeguard, but it’s not a guarantee. Bias seeps in, numbers get taken out of context, and sometimes, an error slips by the expert reviewers.

A story from the book Lies, Damned Lies, and Statistics illustrates the point. A dissertation prospectus once claimed that “every year since 1950, the number of American children gunned down has doubled.” It sounded serious and urgent, but the math does not hold up. If the count started at just one child in 1950 and truly doubled each year, it would have reached more than a million by 1970, well beyond the total number of homicides in the entire country. Continue that pattern into the 1980s and the number would have outgrown the U.S. population itself. It is a clear reminder that dramatic statements are not the same as solid evidence, and even peer-reviewed work can sometimes miss the mark.

Using AI to strengthen research literacy

To earn a Ph.D. in criminal justice, criminology or sociology, it often takes four to six years beyond a master’s degree. Most of us don’t need that level of specialization. What we do need is the ability to find good research and understand what it tells us.

So, the good news is that you don’t need to spend years and years pursuing a Ph.D. to interpret research intelligently. With a structured approach and the right AI tools, anyone with a basic research competence can learn to find and understand research well enough to help them understand best practices in their agency.

Here’s an effective sequence for doing it.

Start broad — Paper Digest

Think of this as your digital literature-review partner. It scans new academic papers and gives you concise summaries across many fields, including policing and corrections. Search a topic, use-of-force policies, community engagement, crisis intervention and it shows which studies are shaping current thinking. It’s the perfect starting point before drilling deeper. PaperDigest.org.

Trace the network — ResearchRabbit

After finding a key article, plug it into this platform. It visually maps which papers cited it and how ideas connect over time. For a department analyst studying de-escalation models, it’s a quick way to spot emerging patterns and see whether findings have been widely accepted by other researchers. Typically, the more other researchers have cited the article, the more relevant the article is. Researchrabbit.ai.

Break down complexity — ExplainPaper

Most academic articles read like a foreign language. Upload one here, and it will interpret sections in easy-to-understand language. You can even select the reading level — undergrad, graduate, or whatever detail you need. Explainpaper.com.

Ask deeper questions — Humata AI

This tool lets you upload a full research paper and ask it direct questions like, “What are the limitations of this study?” or “How was ‘community engagement’ defined here?” It cites exact sections, so you’re not skimming for answers. It helps you engage with the material instead of being overwhelmed by it.

Follow that process, and you’ll understand research with more confidence. It won’t make you a Ph.D. researcher, but it will give you the ability to evaluate evidence, translate findings for your team, and make policy decisions backed by solid data. Humata.ai.

Understanding the limits of data

Even accurate numbers can mislead without context. If traffic enforcement drops, reported crime convictions may look lower, but that might say more about enforcement priorities than actual reduction in crime occurring. The same rules apply to department-generated stats. Even with AI helping interpret data, human judgment is still often needed to determine what those numbers really mean.

Becoming research-literate

By pairing trusted resources like CrimeSolutions.gov with tools such as Paper Digest,
ResearchRabbit, ExplainPaper and Humata AI, you can stay informed about and appreciate the latest evidence without leaving your day job behind to get that Ph.D. This approach keeps policy grounded in actual understanding and not relying on trends or assumptions or the interpretations of others.

The bottom line

Good research remains a helpful way to determine best practices for policing and public safety. AI can help us navigate and understand research faster. However, research is not perfect, so knowing limitations and using human discernment based on real-world experience is also needed to inform us on what is true and valuable. In sum, the information world is not slowing down anytime soon. The question is whether we’ll be able to keep pace and whether we’ll understand what we find.

---

Reference

Wheeler, M. (1976). Lies, damn lies, and statistics: The manipulation of public opinion in America. Liveright.