Give Police New Powers and Tools
Conservative · what the evidence says
An independent, source-checked look at Conservative’s policy “Give Police New Powers and Tools” — what it would actually do across the things that affect your life. Every claim below quotes the source behind it. How this works.
Personal liberty & free speech — Hurts
moderate · moderate confidence
This policy expands police surveillance (facial recognition), suspicion-less or reduced-threshold search powers, and warrantless entry to premises — all of which directly reduce personal privacy and freedom from state coercion. The main caveat is that existing legal safeguards may constrain the worst excesses, but oversight is acknowledged to be lagging the technology.
The evidence
- The policy commits to giving police facial recognition technology and powers to seize knives and track stolen property. — conservatives.com (manifesto) — “giving officers new powers and tools to catch criminals, including technology like facial recognition and powers to seize knives and track down stolen property”
- Live facial recognition creates mass scanning of faces in public — over three million faces scanned in the past year. — pinsentmasons.com (media) — “only 12 false alerts out of more than three million faces scanned in the past year”
- Civil liberties groups argue LFR presents a significant threat to individual privacy and free expression. — theguardian.com (media) — “Civil liberties groups like Liberty and Big Brother Watch argue that LFR presents a significant threat to individual privacy, free expression, and human rights”
- The Biometrics Commissioner has warned that oversight is lagging behind the technology's rapid growth. — theguardian.com (media) — “oversight is lagging behind the technology's rapid growth, and police forces could face legal challenges for misidentifications”
- Scotland's Biometrics Commissioner described the current legal framework as 'patchwork'. — theguardian.com (media) — “described the current legal framework as "patchwork"”
- A majority of the public believe LFR systems represent a move towards a surveillance society. — policinginsight.com (media) — “57% believe the systems are "another step towards turning the UK into a surveillance society"”
- The new warrant-less entry power allows police to enter and search premises without a warrant if they have reasonable grounds to believe electronically located stolen items are present. — jmw.co.uk (media) — “police officers can now enter and search specified premises without a warrant if they have reasonable grounds to believe that stolen items, electronically located (e.g., via GPS, Wi-Fi, Bluetooth, or tracking devices lik…”
- Critics suggest the warrant-less entry power could place burdens on law-abiding citizens and risks scope creep beyond electronically tracked goods. — spectator.com (media) — “such legislation might not deter criminals effectively but could place greater burdens on law-abiding citizens, with a risk of the powers being expanded beyond electronically tracked goods”
- Under existing suspicion-less stop and search (Section 60), Black people were seven times more likely to be searched than white people, indicating expanded search powers carry a disproportionate liberty cost by ethnicity. — theguardian.com (media) — “Under Section 60 (suspicion-less search), Black people were seven times more likely to be searched than white people”
- Existing police use of force affects Black people at roughly 3.5 times the rate of white people. — stop-watch.org (media) — “Black people were subjected to police use of force at a rate approximately 3.5 times higher than white people in England and Wales”
Biggest unknown: Whether Parliament will attach robust statutory safeguards and independent oversight to the new powers, particularly facial recognition, which would limit the liberty impact significantly.
Our reading: O10 scores policies by whether they expand or contract state coercion over individuals' bodies, movements, and privacy. This policy does three things, each of which extends coercive state reach. First, it deploys live facial recognition at scale — scanning millions of faces in public spaces without individual consent or suspicion. Civil liberties organisations identify this as a direct threat to free expression and privacy, and both Biometrics Commissioners have flagged that oversight is lagging the technology in a 'patchwork' legal framework. This is a qualitative expansion of surveillance. Second, knife-seizure powers extend stop-and-search into private premises (bladed articles on private property) and build on existing suspicion-less search powers whose disproportionate impact on Black people is documented at a 7:1 ratio — the liberty cost falls unevenly. Third, the warrant-less entry power to recover electronically tracked stolen property removes a foundational common-law protection (the warrant requirement) for entry to private premises, substituting inspector-level authorisation. Critics flag scope-creep risk beyond the initial framing. Absent the policy, none of these three instruments would be in place; they represent genuine additions to the state's coercive toolkit, not merely codifications of existing practice. The magnitude is moderate rather than major because existing human rights legislation (PACE, HRA 1998, data protection law) provides some floor of constraint, and the facial recognition false-positive rate in controlled conditions is low. But the oversight gap, the documented disproportionate impact of existing search powers, and the warrantless-entry precedent together justify a clear 'worsens' verdict. Confidence is moderate because the real-world operational scope of deployment remains unspecified.
Crime, justice & national security — Helps
moderate · moderate confidence
Giving police facial recognition, expanded knife-seizure powers, and warrant-free stolen-goods searches adds real enforcement tools that have already generated thousands of arrests and seizures. The size of the crime-reduction benefit is uncertain because accuracy and deterrence effects are still contested.
The evidence
- The policy commits to giving police facial recognition technology, powers to seize knives, and powers to track down stolen property, alongside supporting lawful use of force. — conservatives.com (manifesto) — “giving officers new powers and tools to catch criminals, including technology like facial recognition and powers to seize knives and track down stolen property, while supporting lawful use of force”
- The Metropolitan Police reports 1,700 arrests attributed to live facial recognition. — policinginsight.com (media) — “The Metropolitan Police claims that LFR has already led to 1,700 arrests”
- In 2024-25, the Met recorded 962 arrests from 2,077 potential LFR matches with a false positive rate of 0.0003%. — post.parliament.uk (government) — “the Met reported 962 arrests from 2,067 true alerts out of 2,077 potential matches, indicating a false positive rate of 0.0003% for LFR at a 0.6 threshold”
- 100,000 knives and offensive weapons have been removed from streets since 2019, with 40,000 seized through stop and search. — homeofficemedia.blog.gov.uk (government) — “100,000 knives and offensive weapons have been removed from the streets since 2019, with 40,000 seized through stop and search”
- The Crime and Policing Act 2026 has enacted a power allowing police to enter and search premises without a warrant to recover electronically tracked stolen property during a 'golden hour' window. — jmw.co.uk (media) — “police officers can now enter and search specified premises without a warrant if they have reasonable grounds to believe that stolen items, electronically located (e.g., via GPS, Wi-Fi, Bluetooth, or tracking devices lik…”
- Independent assessment found earlier LFR trials had only a 19% success rate, significantly below the Met's claimed 70%, raising questions about net effectiveness. — theguardian.com (media) — “An audit by Pete Fussey from the University of Essex found that the Met's earlier LFR trials in 2020 had a 19% success rate, significantly lower than the Met's claimed 70%”
- The National Physical Laboratory found an 89% chance of identifying a watchlist subject but noted error rates can rise by up to 9.3% in real-world conditions. — post.parliament.uk (government) — “an 89% chance of identifying someone on a watchlist and a 1 in 6,000 chance of incorrect identification on a 10,000-image watchlist, the error rate of LFR can increase by up to 9.3% outside a testing environment due to f…”
- Research finds focused deterrence achieves a 33% reduction in incidents — six times greater than stop and search — suggesting knife-seizure powers may be less effective than alternative approaches. — theguardian.com (media) — “YEF concluded that other tactics, such as focused deterrence, were more effective, achieving a 33% reduction in incidents, six times greater than stop and search”
Biggest unknown: Whether expanded stop-and-search and LFR deployment actually reduce knife crime and theft at population scale, or whether the arrests recorded represent displacement and net additionality remains undemonstrated by independent evaluation.
Our reading: The policy adds three concrete enforcement tools — live facial recognition, expanded knife-seizure powers, and warrant-free stolen-goods recovery — each of which is now or will be law rather than aspiration. That passes the soft-verb/no-deliverable test: real statutory instruments exist. On O5 alone (the protective good), the evidence supports a genuine improvement signal. LFR has generated documented arrests at a very low false-positive rate in recent deployments (962 arrests, 0.0003% false-positive rate in 2024-25), even if earlier trial accuracy was lower. Knife-seizure powers build on existing stop-and-search infrastructure that has removed 100,000 bladed articles since 2019. The stolen-goods tracking power, now enacted via the Crime and Policing Act 2026, creates a 'golden hour' mechanism that did not previously exist and is designed to improve both recovery rates and deterrence. The counterfactual (absent this policy) is continued reliance on existing PACE powers, which already exist but leave gaps — for example, no warrant-free entry for electronically tracked stolen goods. The new powers fill specific gaps. The magnitude is moderated to 'moderate' rather than 'major' because: independent evidence suggests stop-and-search is less effective than alternative deterrence approaches (YEF finding focused deterrence six times more effective); LFR accuracy in real-world conditions is 9.3% lower than testing environments; and no independent evaluation of the SVRO knife pilot has been publicly released. Civil liberties costs (disproportionate impact, surveillance concerns) are excluded from this verdict per O5's scope and belong on O10. On the purely protective dimension, the balance of evidence supports a real, moderate improvement in the tools available to reduce crime.
Equal treatment & democratic rights — Hurts
moderate · moderate confidence
Expanding police powers in facial recognition, suspicion-less searches, and warrantless entry risks worsening equal treatment for Black and minority ethnic people, who already face documented disparities in stop-and-search and use of force, with oversight frameworks acknowledged to be lagging. The scale of harm depends on whether new safeguards are attached in practice.
The evidence
- The policy proposes giving police facial recognition technology, powers to seize knives, and powers to track and recover stolen property. — conservatives.com (manifesto) — “giving officers new powers and tools to catch criminals, including technology like facial recognition and powers to seize knives and track down stolen property”
- Under Section 60 suspicion-less search, Black people were seven times more likely to be searched than white people. — theguardian.com (media) — “Under Section 60 (suspicion-less search), Black people were seven times more likely to be searched than white people”
- Home Office data shows Black people were subjected to police use of force at roughly 3.5 times the rate of white people. — stop-watch.org (media) — “Home Office data shows that Black people were subjected to police use of force at a rate approximately 3.5 times higher than white people in England and Wales in the year ending March 2025”
- The rate of force used against Black men aged 18–34 was 153.3 per 1,000 compared to 43.1 per 1,000 for white men of the same age. — stop-watch.org (media) — “the rate for Black men aged 18-34 was 153.3 per 1,000 people, compared to 43.1 per 1,000 for white men in the same age group”
- Some studies suggest Black men are more likely to be incorrectly matched by facial recognition algorithms, though this is a contested and partially-evidenced concern. — commonslibrary.parliament.uk (government) — “The House of Commons Library notes concerns about racial bias, with some studies showing Black men being more likely to be incorrectly matched by LFR algorithms”
- Both biometrics commissioners have warned that oversight of facial recognition is lagging behind its rapid growth, and police could face legal challenges. — theguardian.com (media) — “Both the Biometrics Commissioner for England and Wales, Professor William Webster, and Scotland's Biometrics Commissioner, Dr Brian Plastow, have warned that oversight is lagging behind the technology's rapid growth, and…”
- One commissioner described the current legal framework for facial recognition as 'patchwork'. — theguardian.com (media) — “described the current legal framework as "patchwork"”
- The new warrantless entry power for tracking stolen goods requires only inspector-rank authorisation, not judicial sign-off. — mobicode.co.uk (media) — “This power requires authorization from an officer of at least inspector rank”
- Critics warn the warrantless tracking-search power could place greater burdens on law-abiding citizens and risk being expanded beyond electronically tracked goods. — spectator.com (media) — “such legislation might not deter criminals effectively but could place greater burdens on law-abiding citizens, with a risk of the powers being expanded beyond electronically tracked goods”
- Police use of force must be compatible with human rights and equalities legislation under existing law. — commonslibrary.parliament.uk (government) — “police use of force must be compatible with human rights and equalities legislation”
Biggest unknown: Whether new statutory safeguards or judicial oversight will be attached to these powers in practice, and whether the racial disparities already documented will narrow or widen under expanded use.
Our reading: O9 concerns equal treatment, anti-discrimination protections, minority protections, and due process. This policy expands three distinct areas of police power, each with a documented or credible route to worsening equal treatment for minority groups. On stop-and-search: existing suspicion-less Section 60 searches already produce a sevenfold racial disparity. Extending knife-seizure powers without new safeguards builds on a framework that demonstrably treats Black people unequally. There is no cited mechanism in the policy that would reduce these disparities. On facial recognition: some studies (noted by the House of Commons Library, though contested) suggest Black men face higher false-match rates. This is a projected rather than settled finding, but it sits alongside documented real-world misidentification cases and the acknowledgement by two independent commissioners that oversight is lagging and the legal framework is 'patchwork'. A wrongful arrest is a concrete due-process harm, and the risk is not demographically neutral given existing force-use disparities. On warrantless entry for stolen goods: judicial pre-authorisation is replaced by inspector-rank sign-off, weakening due process for occupants of searched premises. Critics flag scope-creep risk, though this is a projected concern. Taken together, the policy materially expands policing powers in domains where existing racial disparities are documented and where oversight is acknowledged to be inadequate. These are direct O9 indicators. The direction is 'worsens'; magnitude is moderate because the harms are real but not universal, and some powers include a 'reasonable grounds' threshold that provides partial constraint.