←
Matt Ashby21,237 posts
Matt Ashby avatar

Matt Ashby

@LessCrime

I help people use data to reduce crime. Associate Professor at @UCLCrimeScience. Former police officer.

◷ Joined April 2009▧ 1,647 Photos and videos
Matt Ashby avatar
Matt Ashby@LessCrime·37m

Are there really professors who believe that the souls of students will be "ruined" if they use a large language model to quiz them ahead of a test or to help them understand difficult language in a set reading? I suspect some colleagues may be losing their sense of perspective.

It's cheaper! It's good! It permits actual learning! It's what the students, when pressed, really want! And, even if you care about things like "employability," it'll make people who will do better in every career, because their minds and souls won't be ruined by screens.

Matt Ashby avatar
Matt Ashby@LessCrime·Nov 12

This paper proposes adopting techniques from red teaming (commonly used to understand adversaries in military strategy and cybersecurity) to help crime prevention practitioners “think thief” (something that’s notoriously hard to do effectively). link.springer.com/article/10…

Matt Ashby avatar
Matt Ashby@LessCrime·Nov 11

Police (🏴󠁧󠁢󠁥󠁮󠁧󠁿🏴󠁧󠁢󠁷󠁬󠁳󠁿) use covert surveillance (almost always interception of phone/internet data) more than twice as often in drug-supply cases than in investigations of violence or sex offences, and five times as often as in fraud cases. doi.org/10.1080/10439463.202…

Matt Ashby avatar
Matt Ashby@LessCrime·Nov 10

Police in two agencies (🇺🇸) that dispatch drones to emergencies ahead of/with officers believe they help improve “tactical decisions, … suspect apprehension, evidence collection” and “improve cost efficiency by … eliminating some in-person responses”. doi.org/10.21428/cb6ab371.a9…

Gavin Hales avatar
⟳ Matt Ashby retweeted
Gavin Hales@gmhales·Nov 6

Greater collaboration between police forces has been talked about for more than a decade as a means of delivering efficiencies, but there are rather few examples of it working in practice. If it's your proposed alternative to mergers, what will be different now?

Matt Ashby avatar
Matt Ashby@LessCrime·Nov 6

Across a large sample of cities (🇺🇸), police agencies introducing degree requirements for new officers didn’t lead to any drop in the gender or ethnic diversity of recruits, as some critics of degree requirements had claimed it would. doi.org/10.21428/cb6ab371.11…

Matt Ashby avatar
Matt Ashby@LessCrime·Nov 5

Always worth remembering: a jury has heard all the admissible evidence in a case, and has been advised on the law by a judge. You probably haven’t heard all the evidence, and don’t know the law as well as a judge.

Matt Ashby avatar
Matt Ashby@LessCrime·Nov 4

Better street lighting and related crime prevention through environmental design improvements halved crime on a high street (🇺🇸) and also led to a reduction on surrounding streets (i.e. a diffusion of benefit, rather than displacement of crime). doi.org/10.1007/s11292-026-0…

Michael Merrifield avatar
⟳ Matt Ashby retweeted
Michael Merrifield@AstroMikeMerri·Nov 3

I was strongly opposed to lecture recordings for much the reasons given here until I received a note from a student with extreme dyslexia, for whom being able to replay short segments of a lecture multiple times had revolutionised his learning. He did very well in his degree.

Lecture recordings devalue education by reinforcing the idea that lectures are just textbook-style content, and encourage students to make poor choices by giving them an inferior but easy option.

Matt Ashby avatar
Matt Ashby@LessCrime·Nov 3

In a lab experiment, increased likelihood/severity of punishment reduced student likelihood of using AI to cheat on an assessment. **But** common university policies tested didn’t influence student perceptions of likelihood/severity, so had no effect. doi.org/10.1016/j.chb.2026.1…