Blocked again on your favourite news site? 7 reasons you’re flagged as a bot and 3 ways to fix it

Your screen freezes, a grey wall appears, and a cold line accuses you of being a robot. It keeps happening to many readers.

That blunt page is not a glitch. It is an automated defence, and it is spreading across major news sites as publishers draw a line against scraping, text and data mining, and the flood of non-human traffic hitting their servers every hour.

What that warning really means

When you hit a message stating your behaviour looks automated, the site’s security system has decided your clicks match patterns used by bots. The notice used by one UK tabloid group sets out two points with unusual clarity: automated access is not allowed, and text or data mining of any kind is banned, including uses for artificial intelligence, machine learning, or large language models.

No automated access. No text or data mining. That includes AI, machine learning, and LLM training, as stated in the site’s terms and conditions.

The same page admits people sometimes get caught in the net. It invites genuine readers to contact customer support if they believe the system made a mistake. It also directs companies seeking commercial licences to a dedicated permissions email.

• Locked out and you are a real reader? Email: [email protected]
• Looking for commercial use or a licence? Email: [email protected]

Why you, a real person, may look like a bot

Security tools weigh dozens of signals. One odd signal rarely blocks anyone. Several at once almost always do. Here are common triggers that catch legitimate readers.

• Using a VPN or corporate proxy that shares an IP with heavy users.
• Opening many tabs quickly, then refreshing pages in bursts.
• Blocking cookies, JavaScript, or key tracking scripts needed for verification.
• Browser automation features or extensions that prefetch, autoscroll, or rewrite headers.
• Very fast, regular click timings that look mechanical.
• Requests from a headless browser or unusual device fingerprint.
• Mobile tethering while travelling between cell towers, which can resemble IP hopping.
• Clock or timezone mismatches, which can confuse integrity checks.
• Tab restore after a crash, causing a flurry of near-simultaneous page loads.

Signals systems look for

The bigger picture: publishers push back against scraping and ai training

Major newsrooms now reject unlicensed scraping and text or data mining, especially for AI training. Their notices, robots files, and paywalls form a layered defence. The stance is simple: if a company wants to reuse journalism at scale, it must seek permission or a paid licence.

Publishers are moving from quiet tolerance of scraping to firm, documented refusal backed by automated enforcement.

This shift aligns with legal frameworks that allow rights holders to opt out of mining. Many sites implement machine-readable signals and formal terms to make that refusal clear. Tools then enforce the policy: behaviour that looks like collection at scale gets throttled, challenged, or blocked.

Three quick fixes if you are suddenly locked out

You can often clear a false flag in minutes. These steps resolve most human-versus-bot mix-ups.

• Wait and refresh: take 10–20 minutes, close extra tabs, then reload a single page.
• Reset your network: disable your VPN, or power-cycle your router to obtain a new IP address.
• Restore verification features: turn on cookies and JavaScript, and disable aggressive privacy add-ons for the site.

Still stuck after 20 minutes with cookies and JavaScript enabled? Contact support and include the error text, the time, your IP, and the page address.

What to include when you email support

Support teams respond faster when they have the basic diagnostics. Keep it short and specific.

• The exact error text on the page.
• Your public IP address at the time of the block.
• The page you were trying to read and the time (with timezone).
• Your browser and device model, and whether a VPN or proxy was active.
• Any case or request ID shown in the message.

For the UK tabloid referenced earlier, use [email protected] for reader access issues. For licensing or commercial crawling requests, use [email protected] and describe your use case, timeframes, and proposed safeguards.

For researchers and companies: how to stay on the right side

If you need consistent, structured access, do not scrape first and ask later. Write to the publisher’s permissions address and request terms. Spell out the data you need, the rate, the storage period, and whether you will use it for AI training. Expect specific limits and audit rights.

• Obtain written permission before any automated collection.
• Respect robots directives and rate limits set by the publisher.
• Separate research crawling from production systems; keep identifiable logs.
• Provide a visible contact and reverse DNS for your fetching infrastructure.
• Offer a takedown mechanism and a way to purge data on request.

Some publishers provide licensed feeds or structured packages. These options cost money but reduce legal and technical risk, and they prevent your IPs from being blocked mid-project.

Privacy tools without the lockouts

Readers value privacy extensions, but heavy-handed settings can trigger defences. Consider a lighter profile for news sites you trust. Allow first-party cookies and core scripts, while keeping third-party tracking reduced. If you need a VPN, pick a dedicated IP option that is less likely to be tainted by abuse from other users.

Beware of “unblocker” plugins that promise instant access. Many route traffic through unknown servers, harvest credentials, or inject ads. A clean browser, reasonable speed, and stable network identity beat risky shortcuts.

Glossary and a quick real-world example

Text and data mining: automated analysis of large volumes of articles to extract patterns or facts. Automated access: non-human retrieval of pages, often at high speed. Headless browser: a browser that runs without a visible window, typically used for scripted tasks.

Scenario: a commuter on mobile data opens ten football stories in new tabs, then toggles a VPN to grab signal on a busy train. The device restores tabs at once, sending near-identical requests from two different IPs within seconds. The site flags the pattern, throws the automated-behaviour page, and blocks further requests for a short period. The reader waits 15 minutes, turns off the VPN, enables cookies, and reloads a single article. Access returns. If the block persists, they email [email protected] with the error text and time.

Small adjustments—steady IP, cookies on, fewer simultaneous requests—often mean the difference between smooth reading and a hard stop.

If you run a newsroom or a research lab, simulate user traffic at human speeds and from stable addresses before scaling up. Build a consented pipeline with the publisher rather than relying on grey-area scraping. The cost is clearer, the dataset is cleaner, and your access does not vanish the moment a filter updates.

Concerned about overblocking? You can test your setup by measuring request spacing, JavaScript execution, and cookie persistence across a short session. Make one change at a time, note the outcome, and keep the configuration that yields stable access without sacrificing your privacy goals.