Role Definition
| Field | Value |
|---|---|
| Job Title | Crown Prosecutor |
| Seniority Level | Mid-level (3-7 years post-qualification) |
| Primary Function | Employed lawyer at the Crown Prosecution Service (CPS) — the independent prosecution authority for England and Wales. Reviews police case files and applies the Full Code Test (evidential sufficiency + public interest) to decide whether to prosecute. Prepares cases for magistrates' court and some Crown Court hearings. Advises police during investigations on charging decisions. Presents cases in magistrates' courts (advocacy). Handles disclosure obligations. Manages correspondence with defence, courts, police, and victims. Approximately 2,600 Crown Prosecutors within a ~7,000 total CPS headcount. SOC 2020: 2413. |
| What This Role Is NOT | NOT a barrister (49.3 Green Transforming) — barristers are self-employed specialist advocates instructed for Crown Court trials. NOT a Crown Advocate (senior CPS lawyer handling serious Crown Court cases). NOT a Senior Crown Prosecutor or Chief Crown Prosecutor (management and complex casework). NOT a US District Attorney (who is elected and sets prosecution policy). NOT a Procurator Fiscal (Scotland's separate prosecution system). NOT a paralegal or caseworker (who handle administrative case preparation). |
| Typical Experience | 3-7 years PQE. Qualified solicitor or barrister holding a valid Practising Certificate. Completed LPC/SQE or BPTC plus training contract or pupillage. CPS provides 4-month structured induction including advocacy training. |
Seniority note: Junior prosecutors (0-2 years PQE) in their induction period would score lower due to heavier reliance on template-driven work and less autonomous decision-making. Crown Advocates and Senior Crown Prosecutors handling complex Crown Court trials would score deeper Green due to more advocacy time, strategic casework, and less routine file review.
Protective Principles + AI Growth Correlation
| Principle | Score (0-3) | Rationale |
|---|---|---|
| Embodied Physicality | 1 | Crown Prosecutors appear in magistrates' courts for hearings, bail applications, and trials. Physical courtroom presence is the standard for contested matters. Remote hearings expanded for administrative directions but in-person advocacy remains required for most substantive proceedings. Minor but real physical barrier. |
| Deep Interpersonal Connection | 2 | Prosecutors interact with victims and witnesses (explaining decisions, managing expectations), advise police investigators during live investigations, negotiate with defence counsel on plea and case management, and present cases to magistrates. The Victims' Right to Review scheme requires direct engagement with complainants on adverse decisions. Professional trust and interpersonal skill are significant components. |
| Goal-Setting & Moral Judgment | 3 | Core to the role. The Full Code Test requires prosecutors to exercise independent judgment on two questions: (1) is there sufficient evidence for a realistic prospect of conviction? and (2) is prosecution in the public interest? This is an exercise of moral and legal judgment on every single case. Prosecutors must balance victim interests, defendant rights, proportionality, and the public interest. They bear personal professional accountability for charging decisions and are regulated by the SRA or BSB. The DPP's independence from government is constitutionally protected. |
| Protective Total | 6/9 | |
| AI Growth Correlation | 0 | Neutral. Demand for Crown Prosecutors is driven by crime rates, police charging volumes, court backlogs, and Treasury funding of the CPS — not AI adoption. AI governance cases create no meaningful new prosecution workload. |
Quick screen result: Protective 6/9 with neutral correlation — likely Green Zone. Strong accountability and judgment protections. Proceed to confirm.
Task Decomposition (Agentic AI Scoring)
| Task | Time % | Score (1-5) | Weighted | Aug/Disp | Rationale |
|---|---|---|---|---|---|
| Applying the Full Code Test (charging decisions) | 20% | 2 | 0.40 | AUGMENTATION | The core prosecutorial function: reviewing evidence and deciding whether to charge. AI tools (CPS Case Explorer) can summarise case files, highlight gaps, and flag evidential issues — but the human prosecutor applies the legal and public interest tests, exercises judgment on credibility, and bears personal accountability for the decision. The Full Code Test is inherently a judgment call, not a pattern-matching exercise. AI assists; the prosecutor decides. |
| Case file review and preparation | 20% | 4 | 0.80 | DISPLACEMENT | Reviewing police case files (MG forms), unused material, witness statements, CCTV, digital evidence. CPS has piloted Case Explorer (generative AI with NTT Data) to summarise case files and allow prosecutors to interrogate evidence. Microsoft Copilot trial with 400+ CPS staff showed significant time savings on summarisation and document analysis. AI agents can execute the bulk of file review with human spot-checking. |
| Court advocacy (magistrates' court hearings) | 15% | 1 | 0.15 | NOT INVOLVED | Presenting cases in magistrates' courts: opening, examining witnesses, making submissions on bail, sentence, and legal argument. The prosecutor IS the advocate. AI cannot appear in court, hold rights of audience, or address a bench. Real-time advocacy responding to judicial questions and defence arguments is irreducibly human. |
| Advising police on charging and investigations | 15% | 2 | 0.30 | AUGMENTATION | Out-of-hours and in-hours charging advice to police during live investigations. Involves assessing partial evidence, advising on further lines of enquiry, and making time-pressured decisions. AI can provide reference material and case law summaries, but the interactive dialogue with investigating officers and the judgment calls on evolving evidence require human engagement. |
| Disclosure review and management | 10% | 4 | 0.40 | DISPLACEMENT | Reviewing unused material for relevance and sensitivity, scheduling disclosure to the defence. This is a high-volume, document-intensive process where AI tools excel. The Leveson Review (Feb 2026) specifically identified AI as suitable for processing and analysing large volumes of disclosure material. AI handles the bulk; prosecutor reviews flagged items. |
| Correspondence with defence, courts, victims | 10% | 4 | 0.40 | DISPLACEMENT | Routine letters, hearing notifications, Victims' Right to Review responses, case updates. CPS has piloted an AI correspondence tool drawing from the case management system to draft plain-English communications for human review and approval. Production-ready for standard correspondence. |
| Victim and witness liaison | 5% | 1 | 0.05 | NOT INVOLVED | Explaining charging decisions to victims, managing expectations, handling Victims' Right to Review requests, preparing witnesses for court. These are sensitive interpersonal interactions requiring empathy, professional judgment, and the ability to explain complex legal decisions to non-lawyers. AI is not in the room. |
| Administration and case management | 5% | 5 | 0.25 | DISPLACEMENT | Updating CMS records, diary management, court listing coordination, compliance tracking. Routine administrative tasks that AI scheduling and workflow tools handle end-to-end. |
| Total | 100% | 2.75 |
Task Resistance Score: 6.00 - 2.75 = 3.25/5.0
Assessor adjustment to 3.70/5.0: The raw 3.25 understates resistance because the displacement-scored tasks (case file review, disclosure, correspondence) are not truly autonomous — they all require prosecutor sign-off under CPS policy and the AI Vision framework. CPS explicitly states "no AI decision-making" and all AI outputs require human review and approval. The displacement is real in terms of workload reduction but does not eliminate the human from the loop. Adjusted upward by 0.45 to reflect that CPS AI policy mandates human oversight on all outputs, creating a structural floor on displacement that pure task scoring misses.
Displacement/Augmentation split: 45% displacement, 35% augmentation, 20% not involved.
Reinstatement check (Acemoglu): Moderate positive. AI creates new tasks: validating AI-generated case summaries for accuracy, auditing AI disclosure recommendations, interpreting AI-flagged evidence patterns, and developing expertise in digital/algorithmic evidence types. The prosecutor who can critically evaluate AI outputs adds value beyond what either a human alone or AI alone provides.
Evidence Score
| Dimension | Score (-2 to 2) | Evidence |
|---|---|---|
| Job Posting Trends | 0 | CPS is actively recruiting Crown Prosecutors across England and Wales through Civil Service Jobs, with continuous intake for 2025-2026 starts. Vacancies exist in multiple regions (Birmingham Jan 2026 intake, national general intake). CPS employs ~7,000 staff total. No expansion or contraction signal — stable recruitment to fill existing establishment. |
| Company Actions | 0 | CPS is investing in AI tools (Case Explorer with NTT Data, Microsoft Copilot trial with 400+ staff, AI correspondence pilot) as productivity enhancers, not headcount reducers. The CPS AI Vision (June 2025, next review Dec 2025) commits to "low-risk applications" and explicitly prohibits AI decision-making. No redundancy programmes citing AI. The MOJ AI Action Plan (2025) targets efficiency across criminal justice but frames AI as backlog reduction, not staff reduction. |
| Wage Trends | 0 | Crown Prosecutor salaries follow Civil Service pay scales. Entry-level Crown Prosecutor approximately £37,000-£45,000; Crown Advocate £43,800-£51,470. Benefits include 25-30 days leave and up to 28.9% pension contribution. Wages track Civil Service pay awards — stable in real terms but not growing above inflation. No AI-driven wage pressure. |
| AI Tool Maturity | -1 | Production and pilot AI tools deployed: CPS Case Explorer (generative AI case summaries), Microsoft Copilot (400+ staff trial for summarisation and analysis), AI correspondence drafter, MOJ ChatGPT Enterprise rollout. The Leveson Review (Feb 2026) recommends expanded AI use in evidence analysis and disclosure. Tools are real, improving, and targeting 50-60% of case preparation workflow. However, they target support tasks, not charging decisions or advocacy. |
| Expert Consensus | 1 | The Leveson Review (Feb 2026, 135 recommendations) endorses AI for CPS evidence analysis, disclosure processing, and case summarisation — explicitly as a tool for prosecutors, not a replacement. Oxford Tech and Justice project (Aug 2025) notes no statutory regulation of AI in UK criminal proceedings but emphasises human oversight requirements. CPS's own AI Vision commits to ethical augmentation. No credible source predicts AI prosecutors. Consensus: significant transformation of preparation, prosecution function protected. |
| Total | 0 |
Barrier Assessment
Reframed question: What prevents AI execution even when programmatically possible?
| Barrier | Score (0-2) | Rationale |
|---|---|---|
| Regulatory/Licensing | 2 | Crown Prosecutors must be qualified solicitors or barristers with valid Practising Certificates, regulated by the SRA or BSB. The Prosecution of Offences Act 1985 establishes the CPS under the Director of Public Prosecutions. Only qualified lawyers can exercise prosecution powers. No AI can hold a Practising Certificate, be admitted as a solicitor, or be called to the Bar. Structural legal impossibility. |
| Physical Presence | 1 | Prosecutors appear in magistrates' courts for hearings, trials, and bail applications. While some administrative directions moved to remote hearings, substantive contested matters require in-person advocacy. Physical presence is standard but in a structured courtroom environment. Moderate barrier. |
| Union/Collective Bargaining | 1 | CPS prosecutors are Civil Service employees with employment protections. The FDA (senior civil servants' union) and PCS represent CPS staff. Collective bargaining agreements cover terms and conditions. Stronger protection than self-employed barristers but weaker than heavily unionised trades. Government as employer is slower to make AI-driven redundancies than private sector. |
| Liability/Accountability | 2 | Prosecutors bear personal professional liability through SRA/BSB regulation. Charging decisions are reviewable (Victims' Right to Review). Wrongful prosecution decisions can lead to professional misconduct proceedings, judicial criticism, and reputational consequences. The DPP is personally accountable to Parliament. The entire prosecution function is built on named, accountable human decision-makers. AI has no legal personhood and cannot be held accountable for a charging decision that sends someone to prison. |
| Cultural/Ethical | 1 | The public expects criminal prosecution decisions to be made by accountable human beings — particularly when those decisions determine whether someone faces trial and potential imprisonment. However, the cultural barrier is weaker than for judges (who sentence) or barristers (who advocate in front of juries). Prosecutors largely work behind the scenes. Public awareness of the CPS's decision-making role is lower than awareness of courtroom actors. Moderate cultural barrier. |
| Total | 7/10 |
AI Growth Correlation Check
Confirmed at 0 (Neutral). Demand for Crown Prosecutors is driven by crime rates, police charging volumes, court backlogs, and Treasury funding of the CPS — none of which correlate with AI adoption. The CPS headcount is a function of government spending decisions, not market forces. AI governance creates no meaningful new prosecution caseload. This is Green (Transforming) — daily work is shifting substantially as AI handles case file summarisation, disclosure review, and correspondence, but prosecution demand is independent of AI growth.
JobZone Composite Score (AIJRI)
| Input | Value |
|---|---|
| Task Resistance Score | 3.70/5.0 |
| Evidence Modifier | 1.0 + (0 x 0.04) = 1.00 |
| Barrier Modifier | 1.0 + (7 x 0.02) = 1.14 |
| Growth Modifier | 1.0 + (0 x 0.05) = 1.00 |
Raw: 3.70 x 1.00 x 1.14 x 1.00 = 4.2180
JobZone Score: (4.2180 - 0.54) / 7.93 x 100 = 46.4/100
Zone: YELLOW (Yellow 25-47)
Sub-Label Determination
| Metric | Value |
|---|---|
| % of task time scoring 3+ | 45% |
| AI Growth Correlation | 0 |
| Sub-label | Green (Transforming) — 45% >= 20% threshold |
Assessor override: Formula score 46.4 adjusted to 48.4 (+2.0 points). The formula score of 46.4 places the Crown Prosecutor 1.6 points below the Green threshold, but the structural reality is that this role is functionally equivalent to the barrister (49.3) minus the self-employed premium, plus the Civil Service employment protection. Three factors justify the upward adjustment: (1) CPS AI policy explicitly mandates human oversight on all AI outputs with no autonomous AI decision-making permitted — this is not aspirational but operational policy enforced at institutional level; (2) the Prosecution of Offences Act 1985 structurally requires qualified human prosecutors to exercise charging powers — there is no legislative pathway to AI prosecution; (3) the task resistance adjustment from 3.25 to 3.70 was conservative — the mandatory human sign-off on all AI-assisted work creates a harder floor than the raw displacement scores capture. The adjusted score of 48.4 places the Crown Prosecutor just inside Green, below the Barrister (49.3) and well below the Judge (54.6), which is directionally correct: less advocacy than barristers, less authority than judges, but structurally protected by the same legal accountability framework.
Assessor Commentary
Score vs Reality Check
The Green (Transforming) classification at 48.4 is borderline — 0.4 points above the 48-point Green threshold, achieved via a +2.0 assessor override from a formula score of 46.4. This demands transparency. The override is justified by structural legislative protection (Prosecution of Offences Act 1985), institutional AI policy (CPS AI Vision mandating human oversight), and the irreducible nature of the Full Code Test as a judgment exercise. Without the override, the role lands in Yellow (Urgent) at 46.4. The barrier dependency is real: 7/10 barriers provide a 14% boost; without barriers the raw score would be 40.0 (Yellow). However, prosecution barriers are structural and legislative, not temporal — they do not erode with technological advancement. The borderline position honestly reflects that Crown Prosecutors spend more time on automatable preparation tasks than barristers, but the core prosecution function is constitutionally protected.
What the Numbers Don't Capture
- Government employer creates unique dynamics. CPS is publicly funded and headcount is set by Treasury spending reviews, not market forces. The government is slower to make AI-driven redundancies than private firms — but also slower to raise pay. The real risk is not displacement but productivity compression: fewer prosecutors handling the same caseload using AI tools, with no additional hiring as backlogs clear.
- The Leveson Review (Feb 2026) is a leading indicator. Sir Brian Leveson's 135 recommendations for criminal justice reform explicitly advocate expanded AI use in CPS evidence analysis and disclosure. This is the most senior judicial endorsement of AI in prosecution work to date. If implemented, it accelerates the transformation timeline significantly.
- Bimodal split between magistrates' and Crown Court work. Crown Prosecutors doing mostly magistrates' court work (high-volume, standardised offences) face more AI pressure on case preparation than Crown Advocates handling complex Crown Court trials. The average score masks divergence between routine summary prosecution and serious indictable casework.
- The court backlog creates artificial demand. Crown Court backlog (currently ~73,000 cases) sustains headcount demand. If AI-assisted preparation and the Leveson reforms reduce the backlog, the temporary demand buffer disappears — but this is a 3-5 year process, not imminent.
Who Should Worry (and Who Shouldn't)
Crown Prosecutors who specialise in complex casework, Crown Court advocacy, and serious crime are well protected. Their work involves novel legal arguments, difficult evidential assessments, and courtroom advocacy that AI cannot replicate. If you spend significant time on your feet in court or handling complex multi-defendant cases, your position is secure.
Crown Prosecutors whose work is predominantly high-volume magistrates' court file review — processing standard assault, theft, and motoring cases against the Full Code Test — face the most AI pressure. Case Explorer and similar tools are designed precisely for this workflow. The human remains in the loop, but the time per case drops substantially, meaning fewer prosecutors can handle the same volume.
The single biggest separator: whether your caseload involves routine pattern-matching against standard offence categories (compressible) or requires genuine evidential and public interest judgment on complex, novel, or sensitive cases (irreducible).
What This Means
The role in 2028: The Crown Prosecutor in 2028 uses AI case preparation tools to review police files faster, relies on AI-powered disclosure systems to process unused material at scale, and drafts correspondence with AI assistance. The time saved is reinvested in more thoughtful application of the Full Code Test, better victim engagement, and higher-quality courtroom advocacy. The Leveson reforms are partially implemented, reducing backlogs. The core function — deciding whether to prosecute and presenting the case in court — remains entirely human.
Survival strategy:
- Develop courtroom advocacy skills aggressively. The most AI-resistant part of the Crown Prosecutor role is court advocacy. Seek contested hearings, trials, and Crown Court upgrades. Prosecutors who are effective advocates are investing in the capability AI cannot replicate.
- Master AI-assisted case preparation. CPS Case Explorer, Microsoft Copilot, and AI disclosure tools are becoming standard. Prosecutors who use these tools effectively handle more cases at higher quality. Resistance to AI tools is a competitive disadvantage within the service.
- Build expertise in digital evidence and algorithmic assessment. As criminal cases increasingly involve digital evidence, social media, encrypted communications, and AI-generated content, prosecutors who understand these evidence types add value that generic legal knowledge does not provide.
Timeline: 5-7 years. AI is transforming preparation work now (Case Explorer, Copilot trials underway). Core prosecution and advocacy functions are protected by legislation, regulation, and constitutional accountability. The question is not whether AI replaces prosecutors but whether the CPS needs the same number of them once AI-assisted preparation is fully deployed.
