Why is Generative AI not blocked by default on LGFL ?

[xicor]

Can I ask what you are counting as adult focused or adult content though. You keep saying craiyon is an adult site or contains adult content, but others aren't seeing it. So, could it be what you are considering adult content, others of us aren't as concerned about? Can you describe, or post screenshots, suitably censored if needed? Are we talking nudity? Adult themes (gambling, guns, violence)?

 

I've looked at recent and inspirations and not seen anything that would make me double take. Searching for "boobs" on the search page gets a little closer to the bone, but I would hope your filtering catches easy searches like that as it is in the URL.

 

Here's some examples of adult content or at least what I class as adult content - all I accessed from within lgfl

 

Unfiltered NSFW AI Chat – Interactive AI Characters & AI Companions

 

another example below, LGFL classes it as general

 

spyce.ai

[sigma]

You're being too hard on LGFL. Its a Netsweeper problem not theirs as I also get the same categorisation for both.

 

I've put in a recategorisation request for both, which should fix it for both of us. I find I'm doing this "all the time", for all different sites that should be falling into blocked categories automatically.

 

(If you scroll down to the bottom on the left of spyce.ai, it even has an "essay grader" for teachers!)

[xicor]

You're being too hard on LGFL. Its a Netsweeper problem not theirs as I also get the same categorisation for both.

 

I've put in a recategorisation request for both, which should fix it for both of us. I find I'm doing this "all the time", for all different sites that should be falling into blocked categories automatically.

 

(If you scroll down to the bottom on the left of spyce.ai, it even has an "essay grader" for teachers!)

 

I didn’t see that essay grader that’s crazy!

 

I also asked them to recategorise those sites along with several other sites, I get they use Netsweeper but I don’t see that as an excuse or me being harsh on them. It is appreciated to challenge them to do better as they are the service providers for safe internet and they are big enough to fix this issue, at least that’s how I see it.

 

I spent very little time and identified many sites that are not categorised correctly, I don’t think LGFL should ignore this issue and should hire dedicated staff to specifically search for those things and block or categorise correctly.

[xicor]

We also use Netsweeper, and have been through the same loop and keep going through the same loop. I don't think that Netsweeper is particularly good at managing blocking Pornography. We get so many false positives that have to be checked, yet our ISP's (Not LGFL) "out of the box" configuration does allow access to pornography, even though it's category blocked.

 

The latest DfE Policy Paper I can find is:

 

https://www.gov.uk/government/publications/generative-artificial-intelligence-in-education/generative-artificial-intelligence-ai-in-education#opportunities-for-the-education-sector

 

The bit we are mindful of is:

 

Protecting data, pupils and staff

 

Generative AI:

 

stores and learns from the data it is given – any data entered should not be identifiable

can create believable content, including more credible scam emails requesting payment – people interact with generative AI differently and the content may seem more authoritative and believable

 

Schools and colleges should:

 

protect personal and special category data in accordance with data protection legislation

not allow or cause intellectual property, including pupils’ work, to be used to train generative AI models, without appropriate consent or exemption to copyright

review and strengthen their cyber security by referring to the cyber standards – generative AI could increase the sophistication and credibility of attacks

ensure that children and young people are not accessing or creating harmful or inappropriate content online, including through generative AI - keeping children safe in education provides schools and colleges with information on:

what they need to do to protect pupils and students online

how they can limit children’s exposure to risks from the school’s or college’s IT system

refer to the filtering and monitoring standard to make sure they have the appropriate systems in place

 

Which leads back to:

 

KCSIE, and the four areas of risk:

content: being exposed to illegal, inappropriate, or harmful content, for example:

pornography, fake news, racism, misogyny, self-harm, suicide, anti-Semitism,

radicalisation, and extremism.

contact: being subjected to harmful online interaction with other users; for

example: peer to peer pressure, commercial advertising and adults posing as

children or young adults with the intention to groom or exploit them for sexual,

criminal, financial or other purposes.

conduct: online behaviour that increases the likelihood of, or causes, harm; for

example, making, sending and receiving explicit images (e.g. consensual and non-

consensual sharing of nudes and semi-nudes and/or pornography, sharing other

explicit images and online bullying, and

commerce: risks such as online gambling, inappropriate advertising, phishing and

or financial scams.

 

Just because its generative AI (new and shiny), it doesn’t mean the statutory guidance doesn’t apply with regards to both blocking and over-blocking.

 

Netsweeper classified this as porn last week. It had me wondering if its location aware?

 

[ATTACH=CONFIG]72603[/ATTACH]

 

Oddly, the mag was featured on HIGNFY this week.

 

Have you considered any alternatives to Netsweeper ? Its really weird such dedicated systems are bad at blocking adult sites or bad at wrongly classifying a site as an adult site!

 

If you don’t mind me asking, how much do you pay for Netsweeper ? And do you have a dedicated account manager to raise those issues with ?

[sigma]

I didn’t see that essay grader that’s crazy!

 

I also asked them to recategorise those sites along with several other sites, I get they use Netsweeper but I don’t see that as an excuse or me being harsh on them. It is appreciated to challenge them to do better as they are the service providers for safe internet and they are big enough to fix this issue, at least that’s how I see it.

 

I spent very little time and identified many sites that are not categorised correctly, I don’t think LGFL should ignore this issue and should hire dedicated staff to specifically search for those things and block or categorise correctly.

 

I find similar with 1 to 1 video chat with strangers sites. I don't know why Netsweeper has such a problem categorising them. They all look the flipping same and use the same language and are a real safguarding nightmare.

[Cat_Jam148]

I spent very little time and identified many sites that are not categorised correctly, I don’t think LGFL should ignore this issue and should hire dedicated staff to specifically search for those things and block or categorise correctly.

 

I'm sure LGFL has a team for this but with just how big the internet is, I feel that all providers will fail at this in some form or another. There are millions of pages created each day, with probably thousands of new domains, and categorising each and every one correctly is never going to happen...even with ai! I mean look at how many youtube to mp3 converters pop up every day, we block them here, but there is always a dozen which sneak through for awhile.

 

If you go deep enough into an innocent Google search, you will always find content which might not be suitable for schools. All we can do is hope our keywords, and LGFLs filters, are up to it.

 

At a point the tech can only do so much, if you have a user who is determined to access certain content, they will find a route to it. We can only try our best to prevent it.

[mavhc]

spyce.ai literally says "WARNING this site is for adults only.", in conclusion: our filtering systems are terrible

 

Email your providers and ask what's wrong with their classifiers

Edited November 12, 2024 by mavhc

[sigma]

Have you considered any alternatives to Netsweeper ? Its really weird such dedicated systems are bad at blocking adult sites or bad at wrongly classifying a site as an adult site!

 

If you don’t mind me asking, how much do you pay for Netsweeper ? And do you have a dedicated account manager to raise those issues with ?

 

Like yourselves it is bundled from our Broadband supplier.

 

Our main problem is not Netsweeper as such, but inept confinguration from our Broadband supplier.

[sigma]

spyce.ai literally says "WARNING this site is for adults only.", in conclusion: our filtering systems are terrible

 

Email your providers and ask what's wrong with their classifiers

 

I had issues reacently at a Primary school where it let me logon as a primary pupils, access Andrew Tate's website and have a 1 to 1 chat with one of his people.

 

This was a different filtering product.

 

They all seem to me to have issues.

[xicor]

spyce.ai literally says "WARNING this site is for adults only.", in conclusion: our filtering systems are terrible

 

Email your providers and ask what's wrong with their classifiers

 

its really bad it sees pleasures.ai as a general site too until I raised it with them, the spyce it saw it as a general site too and I have asked for it to be recategorised

[xicor]

I had issues reacently at a Primary school where it let me logon as a primary pupils, access Andrew Tate's website and have a 1 to 1 chat with one of his people.

 

This was a different filtering product.

 

They all seem to me to have issues.

 

is it this site ? cobratate.com I just checked and LGFL classifies it as a general site

[LGfLMarkB]

Hi Xicor,

I know we have a meeting set up for tomorrow morning so hope we can help get to the bottom of this and make sure your site is set up in the best way, plus if any learning for us to take forward into changes or development plans our end we can look into it.

Thanks. Speak tomorrow

Mark

Edited November 12, 2024 by LGfLMarkB

[xicor]

Hi ,

I know we have a meeting set up for tomorrow morning so hope we can help get to the bottom of this and make sure your site is set up in the best way, plus if any learning for us to take forward into changes or development plans our end we can look into it.

Thanks. Speak tomorrow

Mark

Hi Mark - I look forward to the meeting, another website LGFL does not see as an adult site is dlsite.com yet sophos clearly classes it as a porn site so would be good to get an explanation of this

Edited November 12, 2024 by xicor

[Zammo]

Not an LGfL or Netsweeper customer so not defending them in any way but:

 

Based on 2024 data, 48.6 million new websites have been created this year. That equals approximately 177,372 websites per day. About 4.34% of all websites are considered new and were created in 2024

 

https://explodingtopics.com/blog/how-many-websites-on-the-internet

 

If you look for workarounds you will find them. Relying purely on a technical solution is folly.

[xicor]

Not an LGfL or Netsweeper customer so not defending them in any way but:

 

Based on 2024 data, 48.6 million new websites have been created this year. That equals approximately 177,372 websites per day. About 4.34% of all websites are considered new and were created in 2024

 

https://explodingtopics.com/blog/how-many-websites-on-the-internet

 

If you look for workarounds you will find them. Relying purely on a technical solution is folly.

There is nothing to defend as for example dlsite.com was categorised as a general/games site yet sophos has it as an adult/porn site - its been that kind of site for 20+ years even more based on the archive.org cache

 

Edited November 12, 2024 by xicor

[mavhc]

That's why you need a) a better classifier, and b) be transparent about how well it works. What percentage of sites were correctly categorised automatically?

[sigma]

is it this site ? cobratate.com I just checked and LGFL classifies it as a general site

 

That's the one. I have previously asked for it to be reclassified to no avail, but have it in one of own own shared lists now for Netsweeper. Netsweeper did block the 1 to 1 chat aspect.

[sigma]

Part of the issue is the way that the Web Filters are marketed. All seem to have different strengths and weaknesses, and are mostly good tools that help us.

 

When your supplier states:

 

“Our network configuration, together with our hosted services, ensures 100% of your school’s traffic is completely safe before it enters your school"

 

in some recent marketing fodder, you know that its laced with a fair bit of snake oil. (Trade descriptions?)

[xicor]

if anyone here has the same issues as I do with the categorization and are an LGFL customer it might be worth speaking with Mark as I am meeting him tomorrow via teams so from my side I dont mind if you joined the meeting as well ?

 

Mark - it might be good to clarify the issues with the bad categorization and your chats with Netsweeper to anyone else interested ? I don't mind from my side who joins especially if they have the same issues as i do, I have submitted to your support many more websites all having the same issue

[LGfLMarkB]

We are happy to help any customer which is struggling - get in touch via [email protected]

 

Might be best separately though because we'll need to look at your school's settings and configuration and how we can improve that and can't be screen sharing that if others are there, for example.

 

Cheers.

Mark

[TechMonkey]

Part of the issue is the way that the Web Filters are marketed. All seem to have different strengths and weaknesses, and are mostly good tools that help us.

 

When your supplier states:

 

“Our network configuration, together with our hosted services, ensures 100% of your school’s traffic is completely safe before it enters your school"

 

in some recent marketing fodder, you know that its laced with a fair bit of snake oil. (Trade descriptions?)

 

Also, they are marketed as content filters, as they filter web content, but not based on the content. The majority are categorisation or URL filters only.

[sigma]

if anyone here has the same issues as I do with the categorization and are an LGFL customer it might be worth speaking with Mark as I am meeting him tomorrow via teams so from my side I dont mind if you joined the meeting as well ?

 

Mark - it might be good to clarify the issues with the bad categorization and your chats with Netsweeper to anyone else interested ? I don't mind from my side who joins especially if they have the same issues as i do, I have submitted to your support many more websites all having the same issue

 

I'm interested in the respones you get.

 

In spite of asking for a categorization check of spyce.ai, it still isn't categorised into any category that would cuse it to be blocked for our pupils or KS1/2. Can you ask about this please?

 

The terms for the site state

 

WEBSITE CONTENT

 

Through the Website and under the conditions stipulated in these Terms of Service, Company allows the users of the Website to view or submit videos or other material and to host, share and/or publish such submissions. The submissions are made accessible and provided to other users under and subject to the conditions stipulated in these Terms of Service.

 

The Website contains sexually explicit content that may impair physical, mental or moral development of minors. If you are under the age of 18 and/or under the age of majority in the jurisdiction in which you reside or from which you are accessing the Website, you are not permitted to use the Website.

 

Normally (today), Netsweeper categorised a photo of a sunset, and a picture of Fred Flintstone and a cute sheep as pornography, so why are they being so tardy with this?

[xicor]

We are happy to help any customer which is struggling - get in touch via [email protected]

 

Might be best separately though because we'll need to look at your school's settings and configuration and how we can improve that and can't be screen sharing that if others are there, for example.

 

Cheers.

Mark

to be clear Mark, the core issue here is your systems are not categorising websites correctly and this has nothing to do with my specific school setup - I have also raised another website BoredHumans.com - Free AI Tools which does have links to various inappropriate links within its main domain and yet the logic I'm getting from your team is that only specific sub links are blocked and not even for all schools only for those that have https decryption.

 

it makes no sense to not block the site as a whole as it has more than one link within its subdomain that is not appropriate, yet the message I'm getting is that all schools that do not have https decryption can still access it and the dodgy links within it are still available, only schools with https decryption wont be able to access it, a logic I do not understand and hope you can clarify tomorrow.

[Dan2025]

This was a little eye opening, going to have to contact my provider.

 

Had a play on a pupil test profile and gah!

 

If you search Power ai NSFW in google and go on the power users link, there's quite a few options. I'd say about half are blocked on Netsweeper, one or two I seriously expected to be blocked due to the naming.. undressai for instance was accessible and had nsfw pics on my screen!

Edited November 12, 2024 by Danwinchelsea

[xicor]

This was a little eye opening, going to have to contact my provider.

 

Had a play on a pupil test profile and gah!

 

If you search Power ai NSFW in google and go on the power users link, there's quite a few options. I'd say about half are blocked on Netsweeper, one or two I seriously expected to be blocked due to the naming.. undressai for instance was accessible and had nsfw pics on my screen!

 

its too crazy, some websites even link to many sites that are all obviously not appropriate - you'd think the categorisation system would use some logic to check the site before allowing access, its nuts

Edited November 12, 2024 by xicor