Introduce a Digital Bill of Rights and regulate AI
Green · what the evidence says
An independent, source-checked look at Green’s policy “Introduce a Digital Bill of Rights and regulate AI” — 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 — Mixed picture
moderate · low confidence
The policy promises greater public control over personal data and free-speech protections online, which would improve liberty — but independent social-media regulation of the kind already seen risks pushing platforms to over-remove lawful content and undermine encryption. The actual effect depends entirely on the detail of the Bill, which is not specified.
The evidence
- The policy commits to giving the public control over their data via a Digital Bill of Rights. — greenparty.org.uk (manifesto) — “giving the public control over data”
- The policy commits to independent regulation of social media and safeguarding elections. — greenparty.org.uk (manifesto) — “ensuring independent regulation of social media, safeguarding elections”
- The policy commits to a precautionary regulatory approach to AI, aligning with European and global efforts. — greenparty.org.uk (manifesto) — “Push for a precautionary regulatory approach to AI, aligning with European and global efforts”
- Existing independent social-media regulation under the Online Safety Act already exists, with Ofcom as the designated independent regulator. — commonslibrary.parliament.uk (government) — “Ofcom is designated as the independent regulator for this Act.”
- Critics including Global Partners Digital and the EFF warn that online content regulation poses serious risks to freedom of expression and privacy. — gp-digital.org (media) — “Critics, including Global Partners Digital and the Electronic Frontier Foundation (EFF), have raised serious concerns that the Online Safety Act poses risks to freedom of expression and privacy online.”
- The broad scope of content regulation creates incentives for platforms to over-remove lawful content. — gp-digital.org (media) — “There is an incentive for platforms to "play it safe" and over-remove content due to significant penalties and sanctions.”
- Concerns have been raised that content regulation could undermine end-to-end encryption. — eff.org (media) — “Concerns have also been raised about the Bill's potential to undermine end-to-end encryption, with some viewing it as a blueprint for repression.”
- Campaigners warn that related data legislation makes it harder for people to control their personal data and enables automated decisions. — openrightsgroup.org (media) — “campaigners like the Open Rights Group have warned that the Data Use and Access Bill (a separate but related legislative effort) contains provisions that could make it harder for people to control their personal data and…”
- There are concerns that ministers can change how data is shared with minimal parliamentary scrutiny. — openrightsgroup.org (media) — “ministers can change how data is shared with minimal parliamentary scrutiny”
Biggest unknown: Whether the Digital Bill of Rights would genuinely enshrine strong free-expression and privacy protections or follow the Online Safety Act model, which critics argue threatens both.
Our reading: The policy has two faces on O10. On the privacy/data-control side, the stated commitment to 'giving the public control over data' points toward an improvement in one of O10's core indicators — freedom from surveillance and coercion by automated systems. On the free-expression side, however, the commitment to independent social-media regulation sits in a context where that model already exists (the Online Safety Act/Ofcom) and has attracted credible, sourced criticism: the structural incentive for platforms to over-remove lawful content, the risk of forcing general content monitoring, and the potential to undermine end-to-end encryption are all projected harms with named independent critics behind them. The AI precautionary approach, by aligning with the EU's risk-based model rather than the UK's current 'pro-innovation' stance, could expand regulatory constraints on automated decision-making — cutting both ways, protecting individuals from AI-driven coercion (O10 positive) but also potentially adding licensing/compliance burdens (O10 negative). Crucially, the policy text is aspirational: it says 'introduce' and 'push for' but specifies no mechanism, no rights catalogue, no statutory duties, and no enforcement model. We cannot know from the text whether the Bill of Rights would enshrine robust free-expression and privacy floors or simply lend legitimacy to a restrictive content-moderation regime. Both outcomes are credibly supported by the evidence — hence mixed. Confidence is low because the decisive variable (the Bill's actual content) is entirely unspecified.
Prosperity & living standards — Mixed picture
minor · low confidence
A precautionary AI regulatory approach could constrain the UK's AI-driven productivity growth in the near term, but greater regulatory clarity and public trust might support longer-run investment; the policy's soft commitments and lack of specific mechanisms make any real-world effect hard to predict. The biggest risk is that aligning with the EU's stricter framework raises compliance costs for UK AI businesses without clear offsetting gains.
The evidence
- The policy proposes a precautionary regulatory approach to AI, aligning with European and global efforts. — greenparty.org.uk (manifesto) — “Push for a precautionary regulatory approach to AI, aligning with European and global efforts.”
- The UK's current approach to AI regulation is explicitly 'pro-innovation' and principles-based, contrasting with the EU's more prescriptive, risk-based legislative framework. — baringa.com (media) — “The UK has adopted a "pro-innovation" and "principles-based, context-driven" approach to AI regulation, contrasting with the EU's more prescriptive, risk-based legislative framework (the EU AI Act).”
- The UK government's current AI policies have mainly focused on boosting the AI sector, promoting innovation, and driving growth and productivity. — commonslibrary.parliament.uk (government) — “The government's policies have mainly focused on boosting the AI sector, promoting innovation, and driving growth and productivity.”
- Companies may adopt EU AI standards globally to avoid managing multiple competing requirements, limiting the UK's ability to differentiate its regulatory environment. — policyreview.info (media) — “Some analysts suggest the EU's AI Act may exert an "extraterritorial policy influence" on the UK due to the EU's market size and regulatory capacity, potentially leading companies to adopt EU standards globally to avoid …”
- An impact assessment for AI regulation anticipated only 'low' monetised benefits for businesses from increased public trust in AI. — assets.publishing.service.gov.uk (government) — “An impact assessment in March 2023 anticipated "low" monetized benefits for businesses in the form of greater demand for AI products and services due to increased public trust.”
- There is a debate about whether a principles-based approach is sufficient to mitigate AI risks without stifling innovation. — lordslibrary.parliament.uk (government) — “There is a debate about whether the UK's targeted, principles-based approach is sufficient to mitigate a broad range of current and future AI risks without stifling innovation.”
Biggest unknown: Whether a precautionary/EU-aligned AI regulatory framework materially reduces UK AI investment and productivity gains compared to the current pro-innovation approach, or whether regulatory certainty and trust produce compensating economic benefits.
Our reading: The existing UK framework is deliberately pro-innovation (E15, E21), treating AI investment and productivity as key economic levers. This policy pushes in the opposite direction — toward a precautionary, EU-aligned stance. Near-term, a more prescriptive regulatory framework would raise compliance costs for UK AI businesses, potentially dampening the investment and firm-formation the current approach is designed to foster. The Brussels Effect (E30) means UK businesses may face EU-level compliance demands regardless, which could either reduce the cost of aligning or, if the UK gold-plates further, add to it. The anticipated benefits from greater public trust were independently assessed as 'low' in monetised terms (E27). On the Digital Bill of Rights side, social media is already independently regulated via Ofcom under the Online Safety Act (E3, E4), so the additionality for prosperity is limited. The debate over whether precautionary regulation stifles or enables innovation is genuine (E28), and the policy's soft-verb framing ('push for') means no committed instrument is in place, limiting the certainty of any effect. On balance, the evidence slightly favours a mild near-term drag on AI-driven productivity and investment relative to the counterfactual, with uncertain longer-run effects from trust and regulatory clarity — hence 'mixed/minor'.
Equal treatment & democratic rights — Little effect
minor · low confidence
The policy promises to strengthen democratic rights in digital spaces and safeguard elections, but uses only aspirational language with no committed legal mechanism or budget — and much of what it describes already partly exists. Until concrete instruments are specified, meaningful improvement to equal treatment or democratic rights is unlikely.
The evidence
- The policy commits to establishing UK standards for rule of law and democracy in digital spaces, independent social media regulation, election safeguarding, public data control, and a precautionary AI approach. — greenparty.org.uk (manifesto) — “Introduce a Digital Bill of Rights to establish UK standards for rule of law and democracy in digital spaces, ensuring independent regulation of social media, safeguarding elections, and giving the public control over da…”
- Independent regulation of social media already exists via Ofcom under the Online Safety Act 2023. — commonslibrary.parliament.uk (government) — “Ofcom is designated as the independent regulator for this Act.”
- The Online Safety Act 2023 already imposes duties on platforms regarding illegal content and harmful content, including protections for democratic debate. — commonslibrary.parliament.uk (government) — “This includes protecting democratic debate and journalistic user-generated content, and a duty to prevent fraudulent advertising.”
- The UK's current AI approach is already principles-based and pro-innovation rather than precautionary, contrasting with the EU model the policy seeks to align with. — baringa.com (media) — “The UK has adopted a "pro-innovation" and "principles-based, context-driven" approach to AI regulation, contrasting with the EU's more prescriptive, risk-based legislative framework (the EU AI Act).”
- Under the current non-precautionary AI framework, there are concerns that present harms such as AI bias go inadequately addressed, which a precautionary approach could partially mitigate. — ukandeu.ac.uk (academic) — “some argue that an excessive focus here without supporting regulators in tackling present harms (like AI bias) could undermine domestic governance efforts.”
Biggest unknown: Whether a Digital Bill of Rights would add substantive protections beyond the Online Safety Act 2023 and existing frameworks, or remain a restatement of aspirations.
Our reading: The policy's stated goals — rule of law in digital spaces, election safeguarding, data control, and precautionary AI governance — all point in a direction relevant to O9 (democratic rights, equal treatment, due process). However, every commitment is expressed through soft verbs ('introduce', 'push for') with no specified statutory duty, budget, enforcement mechanism, or quantified target. Under the soft-verb/no-deliverable rule, this defaults to negligible absent evidence of a delivered mechanism. Critically, much of what the Digital Bill of Rights would cover is already partially addressed: Ofcom is already the independent social media regulator under the Online Safety Act 2023, which already includes protections for democratic debate. The policy would need to meaningfully exceed this baseline to register an improvement. No evidence unit demonstrates what incremental O9 gain a new Bill would deliver over the OSA. On AI, the precautionary shift toward EU-alignment could address present AI-bias harms (which under the current framework risk going under-regulated), and this is the most credible O9-relevant pathway. But 'push for' a regulatory approach is not the same as legislating one, and the existing framework already formally acknowledges bias and discrimination risks. On data control, the policy aspires to give the public more control, but concurrent legislation may move in the opposite direction, weakening individual data rights — undermining the stated goal. Overall, the direction of aspiration is positive for O9 indicators, but the absence of committed mechanisms, combined with significant existing coverage and countervailing legislative trends, means no material population-scale improvement on equal treatment or democratic rights can be projected from the policy as stated. The verdict is negligible rather than improves, with low confidence given the vague policy text.