Theme analysis

AI-Generated Content Labelling

Built-in sample analysis for the AI-generated content labelling consultation.

View briefing Start new consultation

Showing seeded sample analysis while the project route initializes.

Responses

Stakeholder groups

Theme frequency

Evidence gaps

Top themes

Theme frequency

Transparency and labelling

Submissions call for visible and machine-readable indicators of AI-generated content.

Moderate consensus

Stakeholder groups mentioning it

Civil society, Startup, Platform / technology company

Number of responses

Policy relevance

Supports user awareness, platform accountability, and democratic resilience.

Suggested next step

Assess whether labels should vary by content risk, audience, and distribution channel.

Public trust and accountability

Respondents connect policy legitimacy to oversight, appeal routes, and public clarity.

Insufficient evidence

Stakeholder groups mentioning it

Civil society

Number of responses

Policy relevance

Central to adoption, compliance, and confidence in public-facing AI systems.

Suggested next step

Map accountability duties across deployers, platforms, vendors, and regulators.

Evidence and evaluation

Respondents want stronger empirical support before final policy choices are made.

Moderate consensus

Stakeholder groups mentioning it

Civil society, Startup, Platform / technology company

Number of responses

Policy relevance

Determines whether measures are proportionate, enforceable, and outcome-focused.

Suggested next step

Commission targeted evidence on user comprehension, compliance cost, and harm reduction.

Implementation burden

Stakeholders warn that policy duties may exceed current delivery or compliance capacity.

Insufficient evidence

Stakeholder groups mentioning it

Startup

Number of responses

Policy relevance

Affects feasibility, small-organization impact, procurement, and timing.

Suggested next step

Develop staged requirements, templates, and support for smaller organizations.

Innovation and market effects

Some stakeholders caution that broad rules may chill beneficial innovation or entry.

Insufficient evidence

Stakeholder groups mentioning it

Startup

Number of responses

Policy relevance

Requires balancing rights protection with market development and competition.

Suggested next step

Model compliance costs and consider proportionate duties for small firms.

Safeguards against harm

Submissions emphasize prevention, rapid response, and protections for affected groups.

Insufficient evidence

Stakeholder groups mentioning it

Platform / technology company

Number of responses

Policy relevance

Shapes risk thresholds, enforcement priorities, and public protection duties.

Suggested next step

Prioritize safeguards for high-impact contexts and vulnerable affected groups.

Consensus and disagreement

Areas of consensus

Transparency and labelling, Evidence and evaluation

Areas of disagreement

No strong disagreement identified.

Minority but important concerns

Public trust and accountability, Implementation burden, Innovation and market effects, Safeguards against harm

Parliamentary or media sensitivity

Transparency, misinformation, standards, and small-business burden are sensitive until validated.

Evidence gaps and barriers

Implementation barriers

We support transparency for high-risk content, but blanket labelling creates compliance costs for small firms. Evidence is needed on false positives, user comprehension, and the costs of watermarking open-source outputs.

Suggested safeguards

The policy should include audits, penalties for repeated non-compliance, and evidence on whether labels change user behaviour during elections and emergencies. Safeguard not clearly specified The strongest safeguard is a layered approach: content credentials, user-facing labels, researcher access to aggregate data, and clear escalation for harmful synthetic media.

Public trust implications

Trust depends on prominent labels, machine-readable provenance, audits, and credible escalation.

Themes by stakeholder type

Civil society: Transparency and labelling, Public trust and accountability, Evidence and evaluation | Industry: Transparency and labelling, Evidence and evaluation, Implementation burden, Innovation and market effects | Platforms: Transparency and labelling, Evidence and evaluation, Safeguards against harm

Strongest arguments

Disclosure should be prominent and machine-readable.
The policy should include audits, penalties for repeated non-compliance, and evidence on whether labels change user behaviour during elections and emergencies.
We support transparency for high-risk content, but blanket labelling creates compliance costs for small firms.
Government should adopt interoperable standards, provide open testing tools, and phase obligations by risk.
Evidence is needed on false positives, user comprehension, and the costs of watermarking open-source outputs.

Policy sensitivity

Parliamentary or media sensitivity is highest where labelling rules affect elections, emergencies, platform enforcement, open-source tools, or small-business compliance.