A Multidisciplinary Study of Predictive Artificial Intelligence Technologies in Criminal Justice Systems
- Shared screen with speaker view

16:55
Welcome everyone! We'll be starting in a few minutes. Please share your name, affiliation and interests in the chat

17:26
I am Charles Raab (University of Edinburgh/ Alan Turing Institute/ Co-Chair of Turing’s Data Ethics Group/ Member of Home Office Biometrics and Forensics Ethics Group)

17:47
Ruth Drysdale, Turing AI programme manager

18:09
Mike Wald Southampton/Turing AI & Inclusion

18:18
Hi I'm Jessie from the Turing. I'm usually organising these events but here today as a participant as the topic sounds super interesting

18:38
The session will be recorded and uploaded onto the FTP webpage soon afterwards

18:39
Hello - I’m Samuel Bell, PhD student at University of Cambridge Computer Lab’s ML Group, and Turing Enrichment Student. I’m particularly interested in hearing about synthetic datasets that are biased/unbiased.

19:20
@Samuel - are you a member of the Turing Syn Data IG?

19:32
I’m not, but perhaps I should be!

19:35
Hello, I'm Marion Oswald, Northumbria Uni and the Turing Institute. I also chair the West Midlands PCC/Police data ethics committee and sit on the CDEI advisory board.

19:57
Hi, I’m Shannon Vallor, a Turing Fellow and Prof of AI and Data Ethics at The University of Edinburgh. I may need to rejoin as my laptop audio isn’t working well.

20:34
Hello, Makadaa Henry-Nickie at the Brookings Instituton.

41:14
If police keep being despatched to same location wouldn't you expect some crime types to move to a different location where there is no police presence?

43:29
please provide link for resources.

44:19
Hello team, is there any way that we can contribute to this project?

49:16
Southampton-Crimonology-AICJS-Project/PredPol-Investigation: Version 1 (Version v1.0.1). Zenodo. \url{http://doi.org/10.5281/zenodo.431152}0.

51:47
In what way is this study ‘multidisciplinary’ in either the theories or the research methods?

56:05
Are criminals modelling police behaviour to decide where to despatch criminals?

01:01:07
snake oil?

01:04:59
What about using algorithms to find cybercrime where community methods are more difficult - eg what sources of fraud are growing that need more policing/investigation?

01:07:37
in fairness to the police, the algorithmic/data-driven methods and the community methods are not mutually exclusive e.g. the Thames Valley violence reduction unit

01:08:48
Neighbourhood Watch is a kind of precursor f community-based approaches, and it has pros and cons as well as ethical issues about local-level surveillance between people themselves.

01:17:00
I have to run to another call, thanks Pamela and Age for a wonderful presentation!

01:17:19
Thanks for a great talk!