Your School's Web Filter Isn't Effective for 1:1 Devices That Go Off Campus

Updated: Oct 9


Sure, your current solution is filtering off campus. But is it effective?



Example 1 – Blocking games to keep students focused 

Will you block "Google.com" or "snake"?


Example 2 – Figuring out what students are really doing online


Does your reporting only show URLs?

When you have students that look like this on Google Meet...



...but have another tab open on Netflix.




Example 3 – Filtering “adult” content to be CIPA-compliant


The 2nd-grader did a search for "kissing."

Adult content can traumatize children.

Message to reporter from 2nd-grader's mother.

Example 4 – Using VPNs to bypass your proxy filter





Then nothing gets filtered.






The only way to fix all these problems is to use a real-time AI filter (i.e. Deledao Education). Deledao's real-time AI solution is recognized by EdTech Digest in 2020.



Before we go into the technical details, let’s see how Deledao handles your pain points. 


Example 1 – Blocking games to keep students focused  


Artificial intelligence doesn’t use keyword blacklists. 

The keyword "snake" is NOT blacklisted.


AI understands the page like a human.

No embedded games detected.

Example 2 – Figuring out what students are really doing online


Reporting that shows you where, what, and when students are online.

So you finally know what's going on.

Classroom management features to monitor engagement in physical or online classrooms.

At least they're not on Netflix.

Example 3 – Filtering “adult” content to be CIPA-compliant


Filter adult content as it displays in real time.

YouTube's got nothing on you.

Example 4 – Using VPNs to bypass your proxy filter


Browser-based filters are VPN-proof!


Here's a taste of what real-time AI filtering can do: 


Check out our demo tool here.


We’ll dive into the technical reasons why your current solution isn’t enough to protect your students and keep your school humming. There are two main reasons.


1. You’re filtering at the proxy level


Unfortunately, the sheer amount of data that passes through your proxy means that it is not technically feasible for the filter to see details of what’s on the webpage.


Without the browser rendering the data first, the proxy-level filter only sees individual HTTP objects and cannot see the page as it’s displayed to the student. If your reporting shows only URLs, it’s because that’s all the filter can see at the proxy level.


2. Your filter relies on a database of domains and keywords


Your typical filter relies on a database of pre-categorized domains and keywords to determine if content is blocked. That database may contain a list of 10 million, or even 100 million, domains/keywords and it may be continuously updated, but it simply can’t keep up with the volume of newly created content. 


Example – YouTube


We have 500 hours of video uploaded every single minute to YouTube.

-YouTube CEO Susan Wojcicki on CBS 60 Minutes


Websites often expand their content offerings. Content may change dynamically within a website and oftentimes within the webpage itself. The ever-changing internet of today requires schools to accurately filter what's being shown to the students in their browsers. 


This is where Deledao’s real-time AI filtering shines. Because it sees what students see on the page, analysis is more accurate and reporting gets into granular details.


Example – Is AARP a game site? 


When would you think to block AARP?


Example – Social Media


Some websites have endless scroll and are constantly updated with user-generated content.

There's also Twitter, Instagram, Pinterest, etc.

You may be thinking:

My web filter already uses AI or machine learning. It is already deployed at the endpoint, so my reporting is suitable for my purposes. Why switch to Deledao Education?

Because our logo is pretty.

There’s a lot of buzzwords being thrown around out there. It seems that everyone is jumping on the bandwagon of “using AI” or “using machine learning” and filtering in “real time”. You’ll need to do a little digging (and testing!) to figure out the nuances of what each vendor means.


Consider the following:

Just because a filter is filtering in real time and using AI, it is not the same thing as using an AI to filter in real time. 

Are you using an “AI filter” or an “AI database”?


Is your “AI-powered” or “AI-based” filter an “AI database”? An AI database is using AI or machine learning to pre-categorize websites and keywords to compile a database. It runs into the same problems as the typical database described above. It cannot comb through all the new content uploaded to the web and doesn’t account for the various ways that content can load on each webpage.


Regardless if it's an endpoint solution or not, if your "AI" is filtering at the proxy level, it is using an "AI database." Some endpoint solutions simply pass information back to a proxy-level filter. It is not technically feasible to have an AI filter at the proxy level due to the amount of data that passes through at that bottleneck. Claims that AI is being used at that level probably refer to an AI database.


Are you truly filtering in “real time”? 


Let's say your current solution filters in "real time" by continuously scanning the page. However, without AI, how can your filter accurately understand what's happening on the page? Without AI, humans have to do the work of continuously updating rules for the filter to understand the context of a page. Humans can not do this work as quickly as AI.


For example, a filter has the rule of blocking the word "adult" to block mature content. Sure, you are blocking "adult movies," but you're also blocking "adult education." Because you don't want to over-filter, a human (maybe that's you) will have to set up an exception for the term "adult education." But what about "adult size bikes" and "adult vitamins"? There are plenty of more terms that need to be entered as exceptions.


The conclusion is simple. AI and real-time filtering go together. AI filtering must run at the browser level to achieve real-time filtering. You cannot harness the full capabilities of AI without using AI in real time.


Does your “real-time AI filter” run on all your different devices and operating systems?


If you’re paying for “real-time AI filtering,” why should it only work on some devices but not others?


Solutions on iPads are particularly difficult to achieve. Some solutions utilize a combination of static blacklists installed on the school device and PAC (Proxy Auto Config) files to direct some traffic to a cloud proxy for filtering. In this case, all the problems with proxy filtering still apply. 

The best way to sift through buzzwords is to test out the solution for yourself. It's time for you to try out Deledao Education.

Deledao Education is an AI filter that screens through each webpage in real time to analyze words, images, and videos in context. It works on iPads, Chromebooks, Windows, Mac OS, and tablets. 


Good news for you - to help with school closures, Deledao Education is free for the current school year.


Get a demo today!


And follow us on Twitter @DeledaoHQ

160 views