Require human review of workplace AI

Regulatory baseline

The central case moves beyond the UK white paper’s principles-led approach by creating workplace-specific duties and human review rights.

• Public bodies must audit AI systems.
• Employers face compliance duties.
• Productivity delay is the main risk.

Core trade-offs

Workers gain protection from opaque automated decisions. Employers lose some speed and automation savings. If broad, the policy can slow useful AI adoption.

• Workers gain contestability.
• Employers face compliance costs.
• AI productivity may slow.

Illustrative fiscal impact

+GBP 0.2bn to +GBP 4.0bn. Central estimate: +GBP 1.0bn.

• Positive numbers mean public-finance pressure; negative numbers mean Exchequer savings.
• Gross costs are separated from tax, NI and benefit offsets.
• Private business costs are not automatically fiscal costs.
• Behavioural responses widen the range materially.
• This is not an official costing.

Economic impact by 2027-28

• Jobs: May protect some workers, but slower AI adoption can preserve inefficient tasks.
• Wages: Protects against unfair wage and discipline decisions, not a general pay rise.
• Prices: Compliance costs may pass through in AI-intensive services.
• GDP / productivity: Likely negative if rules delay low-risk AI productivity gains.

Assessment

The policy is easier to justify for hiring, discipline and surveillance than for all workplace AI. A broad human-review rule could protect workers while slowing productivity-enhancing adoption.

Confidence: Low. Compliance cost and lost-productivity channels are not officially costed.

Main risks

• Overbreadth: Covering low-risk AI could delay useful productivity tools.
• Regulatory capacity: Existing regulators may lack technical resources.
• Box-ticking: Human review may become formal rather than meaningful.

Safeguards

• Limit hard duties to high-risk uses.
• Fund technical regulator capacity.
• Require audit trails, not blanket bans.

Academic evidence

UK government evidence

Sources

• PolicyLens illustrative scenario methodology for require human review of workplace ai Internal - PolicyLens, 2026
• A pro-innovation approach to AI regulation UK government white paper - Department for Science, Innovation and Technology, 2023
• Artificial intelligence and employment law Parliamentary briefing - House of Commons Library, 2023
• Using AI in the workplace International evidence - OECD, 2024
• Robots and Jobs: Evidence from US Labor Markets Academic article - Acemoglu and Restrepo, Journal of Political Economy, 2020
• Artificial Intelligence and Jobs: Evidence from Online Vacancies Academic article - Acemoglu, Autor, Hazell and Restrepo, NBER, 2022
• Workers' Charter 2026 Party policy source - Green Party of England and Wales, 2026

Other Green policies

• Raise the minimum wage to £15
• Index public-sector pay to inflation
• Add public-pay restoration premiums
• Raise low-paid public-sector wages
• Cap organisation pay ratios at 10:1
• Protect pay in transition schemes
• Create a single worker status
• Equal terms for outsourced workers