REFNATION
JusticeNationalClosed · Final

Should police use AI to review and summarise digital evidence in criminal investigations?

Yes 51%No 49%112 votes cast

Current 1996 disclosure guidance requires police officers to manually review and summarise every potentially relevant file. Modern cases can involve millions of documents from phones, emails, messages and cloud storage. The Home Office has accepted recommendations to legislate for AI use and will pilot tools in 2027 with a view to national rollout.

SourceGOV.UK
Jump to opinions· 3

The most striking thing about this result is how little it resembled a generational culture war over technology. The 55-64s, old enough to remember disclosure done entirely by hand, were the most enthusiastic backers of automating it, while the 45-54s dug in hardest against; there was no clean young-versus-old story here, just a scatter that suggests views on AI in policing turn less on comfort with technology than on trust in institutions to use it well.

That absence of a tidy divide is itself the story: this is a question about process, not principle, and the chamber's near-deadlock echoes the wider unease that has followed the Home Office's own tentative embrace of AI disclosure tools, due to be piloted in 2027 ahead of any national rollout. Britain, on this evidence, isn't hostile to the technology so much as unconvinced anyone has yet answered for its errors.

That tension ran straight through the floor's best arguments. The top Yes voice conceded the case only "so long as they're accountable for it and actually check it's findings", framing support as conditional rather than enthusiastic, while the leading No voice cut to the same worry from the other side, asking simply "who will take responsibility for potential mistake." Neither side, in other words, disputed that AI could speed up the caseload — the argument was over who answers when it gets something wrong.

A two-point margin on a question of accountability, not capability, is about as close as a chamber can get to saying: convince us who carries the blame, and we might be persuaded.

LeftCentreRight
  • The Cultural DisciplinarianHard Right
    YES
  • K
    The Wealth-Tax EgalitarianLeft
    NO
  • The Guarded ProgressiveCentre Left
    YES
  • r08yn· 165
    The DiplomatLeft
    NO
  • MikeHowes· 246
    The Borderless Constitutional HereticHard Left
    NO
  • The Constitutional ModerateCentre Left
    NO
  • CDRook· 181
    The Rights-State ArchitectHard Left
    NO
  • C
    Cadbold· 18
    The Union GuardianCentre Right
    YES
  • G
    GenLee· 290
    The Centre-Right Constitutional DissenterCentre Right
    YES
  • hywel· 20
    Left
    YES
  • H
    Harisa· 81
    The Selective LibertarianHard Right
    NO
  • Jam· 14
    Left
    NO
  • R
    Ricklor· 112
    The Green ConstitutionalistCentre Left
    YES
  • R
    Revere· 373
    The PolymathHard Right
    NO
  • The Competence ConservativeRight
    YES
  • The Pragmatic LevellerLeft
    NO
  • A
    Andymac· 417
    The SageLeft
    YES
  • The Clemency ConservativeRight
    YES
  • D
    drayton· 177
    The Sovereignty EnforcerRight
    NO
  • ThisisKaj· 379
    Leader of PeopleCentre Left
    NO
51%
Yes · 57 votes
49%
No · 55 votes
Yes No
By gender
Male
49/51
Female
55/45
By age
16–24
33/67
25–34
53/47
35–44
44/56
45–54
36/64
55–64
68/32
65+
50/50
By political leaning
H. Left
20/80
Left
33/67
C. Left
55/45
Centre
50/50
C. Right
64/36
Right
65/35
H. Right
33/67
Spread the wordShare
Debate

3 Opinions

Sign in and vote to share your opinion.
CDRook· 181
The Rights-State Architect
Voted noHard Left

Putting sensitive information into AI which it can learn from could aid criminals in the future in how to get away with other crimes.

The Hardline Egalitarian
Voted yesLeft

So long as they're accountable for it and actually check it's findings, I think it's a good way to speed up the caseload processing

Slawomir· 565
The Self-Reliance Disciplinarian
Voted noRight

No, who will take responsibility for potential mistake.