Why AI is blocked at work but fine at home
A work network checks every site against a policy before it loads, and home Wi-Fi has no such policy. AI chats are often blocked at work because the filter puts them in a restricted category or has never seen them. The site is not broken. Your network is following different rules.
If ChatGPT works at home but not at work, nothing is wrong with your laptop or with the site. A managed network makes a decision about each address before the page loads. This article explains how that decision is made and what each kind of block looks like from your seat.
Read next
On this page
Filters sort sites into categories
Most office filters use a large database that labels sites by type: news, shopping, email, file sharing and so on. The administrator chooses which types are allowed, limited or denied. Many filters now include a generative AI category, and many organizations set it to block, or to allow with limits, while a policy is written. So "why is ChatGPT blocked at work?" often has a plain answer. It sits in a category your employer denied, and nobody singled you out.
Why one AI chat opens and another does not
A filter must label a site before it can apply a rule, and new sites appear faster than any database updates. A site the filter has never seen is often "uncategorized", and many networks block uncategorized sites by default until someone reviews them. That is how one AI chat loads while a newer one stays shut. Networks also differ in design. An allowlist permits only named sites. A blocklist permits everything except named sites. On a blocklist network, an AI that is not blocked may simply not be listed yet.
Where a network blocks an AI chat, and what it looks like
The block can sit at three points, and each looks different. A DNS block answers your device's lookup with the address of a notice page or with no usable answer, so you often get a block page or an error at once. A web proxy sits between you and the site. It may show its own notice, or hold the request until it times out. A firewall usually drops the connection silently, which looks like a page that spins for a long time and then fails. Tools vary, so treat this as a guide, not a rule.
A certificate warning is a stop sign
Some organizations inspect secure traffic. To do that on managed devices, they install their own certificate. If your device does not trust it, your browser may show a certificate warning on a site that is normally fine. Do not click through it, and do not accept it to see the page. Close the tab and tell IT what you saw. Also never install an "unblocker" extension or app. Extensions can read what you type and every page you open, and on a work device they can break policy and put your accounts at risk.
Why it works on home Wi-Fi, and why blocks come and go
Your home router has no category list and no allowlist, so it passes the request through. Mobile data belongs to your carrier, which applies different rules. The same page loading there only shows that the policy is local to your work network. It is not a fault in the site, and you did nothing wrong. Blocks also change. Vendors update category databases, administrators edit policies, and a site can be reclassified overnight. A chat that worked on Monday and fails on Thursday is usually a policy change.
What you can do
Write down the exact words on the screen, or take a screenshot, and note the time. Work out which kind of block you have with the guides linked below. Look for an AI or acceptable-use policy. Then ask IT. A request that names one task, the data involved and the tool gets answered faster than "please unblock ChatGPT". The article on how to ask IT to allow an AI tool covers the wording. Until you have an answer, keep work data out of any tool that is not approved.
What this site does and does not do
nottrack.com offers a free AI chat with no account, no sign-up, no card and no daily message limit, and there is an Android app on the download page. This site is not a proxy and cannot tunnel through someone else's network filter. If your network blocks it, that is the network's decision, and no web page can override it. Ask IT before using any AI tool on a work network.
Straight answers
Why is ChatGPT blocked at work but not at home?
Your work network applies a filter policy. It may put AI chat in a blocked category, or block sites it has not reviewed. Home Wi-Fi has no such policy, so the same page loads. The site is identical in both places.
Can an AI chat be blocked by a firewall without any message?
Yes. A firewall often drops the connection silently, so the page spins and then fails. A DNS filter or proxy more often shows a block page. Either way, the cause is the network, not your device.
Why is one AI chat not blocked when others are?
Filters label sites before applying rules. A site missing from the database may be uncategorized, and networks treat those differently. Some allow them, others block them until reviewed. It says nothing about the quality of the tool.
Is it safe to click through a certificate warning to open an AI chat?
No. The warning means your browser does not trust the connection. On a work network it may come from the company's own filter, but you cannot tell that from the page. Stop and ask IT.
Why did an AI chat stop working at work after months?
Usually a policy or category update. Vendors reclassify sites and administrators tighten rules. Note the message and the time, and ask IT whether something changed.
Check the policy, ask IT, and keep work data out of tools that are not approved.
Open unblocked chat