TL;DR
- On March 18, 2026, the UK government will release a report on AI copyright and economic impact as a legal obligation.
- Out of 11,500 consultation responses, 88% supported "Mandatory Full License for AI Learning (Option 1)."
- The government's preferred opt-out model (Option 3) had a support rate of 3%, and the government withdrew its stance.
- While this report does not immediately lead to legal amendments, it directly influences the global direction of AI learning data regulation, including South Korea.
- Web crawling companies must now immediately refine their crawler User-Agent identification system and purpose-based classification system.
March 18, 2026, the UK government will announce a report on AI and copyright, along with an economic impact assessment. This is not just a policy proposal but a legal obligation under the Data (Use and Access) Act 2025 Sections 135/136. If you are a web crawling/data collection company involved in price monitoring, market analysis, AI learning data collection, you should now understand how this report will impact your operations.
What is this report and why is it important now?
The UK AI copyright report is the result of formal consultations conducted from December 2024 to February 2025 on 50 questions. Analyzing over 11,500 responses, this report is a key document determining the future direction of UK copyright law amendments.
| Category | Content |
|---|---|
| Release Date | March 18, 2026 |
| Legal Basis | Data (Use and Access) Act 2025, Sections 135/136 |
| Lead Departments | DSIT + DCMS + IPO (Intellectual Property Office) |
| Number of Responses | Over 11,500 |
| Report Scope | Technical standards, data access, transparency obligations, licensing, enforcement, AI outputs |
This report, along with the FTC AI Policy Statement (March 11) and the EU AI Act High-Risk AI Regulations (August), forms the three major AI regulatory events in the first half of 2026. Also, check out the analysis of the FTC AI Policy Statement for more information.
What were the results of the four policy options?
The consultation results are clear. Out of 11,500 responses, 88% supported the strongest regulation, Option 1 (Mandatory Full License).
Option 0: Maintain Current Law
- Support Rate: 7%
- Impact on Web Crawling: Maintains the current "gray area." Legal uncertainty persists as the determination of legality relies on precedents.
Option 1: Mandatory Full License
- Support Rate: 88% (Overwhelming majority)
- Supporters: Creative industries, publishers, music/film industry as a whole
- Impact on Web Crawling: If using crawling data for AI learning, individual license negotiations become mandatory. Operating costs are expected to increase significantly.
Option 2: Broad TDM Exemption
- Support Rate: 0.5%
- Impact on Web Crawling: Almost unlimited use of crawling data for AI learning. Realistically, the lowest likelihood of adoption.
Option 3: Opt-Out TDM Exemption (Government Original, Withdrawn)
- Support Rate: 3%
- Situation: Initially preferred by the government but faced overwhelming opposition, leading to a "reset" declared in January 2026 and withdrawal of the stance.
- Impact on Web Crawling: Compliance costs for technical opt-out mechanisms like robots.txt were a key issue.
| Option | Support Rate | Impact on Crawling Companies |
|---|---|---|
| Option 0 (Maintain Current Law) | 7% | Maintains uncertainty |
| Option 1 (Mandatory Full License) | 88% | Significant increase in licensing costs |
| Option 2 (Broad Exemption) | 0.5% | Almost no restrictions |
| Option 3 (Opt-Out Exemption) | 3% | Government withdrawal |
The direction presented in the March 18 report is the most critical point to watch.
Why can't AI crawling be blocked with robots.txt?
robots.txt has fundamental limitations as a means to selectively block crawling for AI learning purposes. This is a key reason for the deadlock in the Option 3 discussion.
| Limitation | Explanation |
|---|---|
| Site-Level Control Only | Unable to control individual works (articles, images, videos) |
| Inability to Distinguish Purposes | Cannot differentiate between search indexing and AI learning crawling |
| Lack of Legal Enforcement | Relies on voluntary compliance. No legal sanctions for non-compliance |
| Hostage Effect on Search Traffic | Blocking AI crawlers risks decreasing search exposure as well |
According to Cloudflare data, Googlebot accessed about 8% of sampled URLs over two months, while during the same period, PerplexityBot accessed 167 times more pages and CCBot accessed 714 times more pages. Website operators are hesitant to block Googlebot for fear of losing search traffic.
In January 2026, the UK Competition and Markets Authority (CMA) demanded "meaningful and effective" control over content usage from Google. Cloudflare argued that "it is technically possible to separate Googlebot for search and AI purposes" and urged for a more fundamental solution.
The discussed technical mechanisms as alternatives are as follows:
| Mechanism | Status | Limitations |
|---|---|---|
| robots.txt | Most common | Unable to control individual works |
| ai.txt | New proposal | Standardization incomplete |
| TDM Reservation Protocol | EU "Gold Standard" | Low adoption rate |
| Meta Tags (noai, noimageai) | Increasing use | Requires voluntary compliance by crawlers |
Where does the UK stand compared to the EU, South Korea, and the US?
The UK has not yet chosen among the four options. This is the biggest difference from the EU and Japan.
| Country/Region | Current Approach | TDM Exemption Status | Key Deadline |
|---|---|---|---|
| UK | Report under preparation | Undecided (Evaluation of 4 options) | Report on 3/18 |
| EU | AI Act + DSM Directive | Art.4 TDM Exemption (Opt-Out) | Full Enforcement in August |
| South Korea | AI Basic Act Enforced | Undecided (Under discussion) | Enforced on 1/22 |
| US | FTC Regulation + Case Law | None (Relies on Fair Use) | FTC Statement on 3/11 |
| Japan | Broad TDM Exemption | Exemption for Information Analysis Purposes | Stable |
| Singapore | Broad TDM Exemption | Commercial TDM Exemption | Stable |
The EU has already confirmed the opt-out-based TDM exemption in the DSM Directive Article 4. However, in the UK consultation, the fact that 88% supported a stronger regulation (mandatory full license) than the EU is noteworthy. There is a possibility that the UK may choose stricter regulation than the EU.
In the case of South Korea, the AI Basic Act was enforced as the world's second comprehensive AI law on January 22, 2026. However, the scope of TDM exemption has not been finalized. The Ministry of Culture, Sports, and Tourism is pushing for legislation, and there is a high possibility that the conclusions of the UK report will be used as a reference for Korean legislation. You can find detailed regulations for each country in the 2026 AI Regulation and Web Crawling Global Guide.
Expected Timeline and Actions to Take Now After 3/18
What will be announced on March 18, 2026, is the report and economic impact assessment. It does not mean that laws will change immediately.
| Timeline | Expected Events |
|---|---|
| March 18, 2026 | Announcement of the report + economic impact assessment |
| Spring 2026 | Proposal of policy direction paper (white paper or draft bill) |
| Summer 2026 | Incorporation of pilot results from the Creative Content Exchange |
| Late 2026 to 2027 | Parliamentary review of copyright law amendments (expected) |
| 2027-2028 | Implementation of amended laws (expected) |
The three key signals to watch for in the March 18 report are:
- Explicitly stating the "Preferred Option" — If specified, it accelerates legislation; if not, it leads to additional consultations and ongoing uncertainty.
- Specific Level of Transparency Obligations — Whether it is at the level of "disclosing learning data summaries" or "individual work lists."
- Recommendations from the Technical Standards Working Group — Whether alternative or complementary mechanisms to robots.txt are proposed.
The actions that web data collection companies need to take now are clear:
Phase 1 — Immediate (Before 3/18)
- Check the current crawler's User-Agent identification system
- Document purpose-based classification for crawling
- Confirm UK business or data collection status
- Appoint a monitor for the 3/18 report
Phase 2 — After the Report Release (3/18~4/30)
- Identify scenarios relevant to your services
- Conduct internal briefings and seek legal advice if necessary
Phase 3 — When Policies are Specified (Second Half of 2026)
- Design a compliance system for transparency obligations
- Simulate licensing costs if Option 1 is adopted
- Establish a system for opt-out detection/compliance if Option 3 is adopted
FAQ
Q. Will the law change immediately after the 3/18 report is released?
No. The report is a mandatory publication under the Data (Use and Access) Act Section 135/136, providing guidance on legislative direction. Actual legal amendments require further consultations and parliamentary review, with an expected lag of at least 1-2 years.
Q. Will Korean companies also be affected by the UK report?
They will be directly affected if they collect data from the UK or provide services to UK customers. Additionally, the conclusions of the UK report are likely to be used as reference material for ongoing TDM exemption legislation in Korea.
Q. If we comply well with robots.txt, will there be no legal issues?
Currently, compliance with robots.txt does not guarantee the legality of crawling. Especially for crawling for AI learning purposes, separate legal evaluations are necessary. The UK government's Technical Standards Working Group officially recognizes the limitations of robots.txt and is considering alternatives. Check the Web Crawling Legality Guide for detailed legal criteria.
Q. How is the TDM regulation in the EU AI Act different?
The EU has already confirmed the opt-out-based TDM exemption in the DSM Directive Article 4. The UK has not yet chosen among the four options. Particularly noteworthy is that 88% of respondents in the UK supported a stronger regulation (mandatory full license) than the EU. There is a possibility that the UK may choose stricter regulation than the EU.
Q. Will crawling for price monitoring or market analysis purposes also be regulated?
The current discussions mainly focus on data collection for AI learning purposes. Traditional business intelligence purposes such as price monitoring and competitor analysis are likely not direct targets of AI learning regulations. However, transparency obligations and crawler User-Agent identification systems can impact all crawling services regardless of purpose.
The UK AI copyright report is a directional guide rather than a law, but with 88% supporting a mandatory full license and the government retracting its original stance, the extent of its impact is hard to predict. In this time of uncertainty, now is the best time to refine purpose-based classification systems for crawling and build trust as a "legal data supplier."
If you are curious about legal operating methods for web data collection, request a free 30-minute consultation with Hashscraper. With over 8 years of experience operating legal data collection services from over 5,000 sites since 2018, we will share our expertise with you.
References
- Copyright and AI: Statement of Progress — GOV.UK
- AI and Copyright: UK Outlook for 2026 — Hogan Lovells
- Government Releases Progress Report — Burges Salmon
- Google's AI Advantage — Cloudflare Blog
- AI and Copyright: Post-Data Bill Timeline — Pinsent Masons
- Creative Content Exchange and UK AI Policy — Pinsent Masons
- South Korea: Comprehensive AI Legal Framework — Library of Congress
