We developed technology to fight back
 

Over 8 months of development, and we are ready to launch some technology that hopefully will be the pitbull our industry needs to reduce the money that is made by unethical webmasters.

We also understand sponsors don't always know where their traffic is coming from, and we can locate your banners and scrub the site for alerts etc.

Its complex, but do you think this will help our present problems?

Sniffy our pitbull was developed to sniff websites and do a deep scrub of affliliate programs, link lists, search engines etc. He checks everything as well as downloads all the images, hashes them and records the fingerprint in a database.

We are able to scrub sponsor programs based on any triggers they want or we use the triggers we know of [lolita, underage, young etc].

The ASACP and our technology have come to an agreement to work together since this technology has so many applications.

Sponsors, do you think this type of service has value to you to have a 3rd party agency sniffing your affiliates and policing their websites on a monthly fee basis?

We have other modules that cross reference this database, and we have taken on the huge task of downloading every image we can find on the net that is adult, to date we have over 5 million in our database.

Our solution to combatting unethical content/websites, is both a high tech and a low tech solution. The low tech side is a human reviewing the images on the webmasters site, as our system recognizes the images we never have to validate them again, so over time we will have less and less to view.

If a webmaster turns on some unethical content during certain times of the week and reverts it back to legit content, we know about it.

Nothing like this has been developed and we feel tools should be developed within our industry.

Any thoughts? Does this have value? If you were a sponsor would you pay a monthly feel for this police dog, we loving call "Sniffy"?

Do sponsors see there is a problem with their affiliates potentially being a CP site, and therefore their sponsor dollars goes to support it?

Well?

That's pretty interesting. Just hope legitimate webmasters don't get caught up in the automated process!

i would be interested in the application towards stolen content.

could it sniff out stolen content, say even if it was resized or cropped?

Since no one else seems to know what to think, I'll gladly reply... based on just the facts you've given, Lee, this sounds like a fantastic solution to start ridding us of the cheaters.
I have some questions, though.
Is this something that you worked up through your old content avenues etc. or is this a Fantasyman project? (Just curious)
If Sniffy wrongly accuses someone as verified by your human eye low-tech factor, will "you" (in whatever capacity the "we" is in your statements) also stand as mediaries between the webmaster and the sponsor program who now has him labelled as a cheater and child pornographer?
When Sniffy flags someone for suspect content/keywords etc., is the human eye involved BEFORE reporting it to the sponsor, or does Sniffy do this automatically on "his" own?

Legitimate webmasters cannot be harmed since we have human reviewers to eyeball the images, and of course any finding Sniffy finds we validate before reporting to sponsor, SE, linklist etc.

If it can find stolen content definetly worth proceeding.

so basically, your technology is looking for words like "underage", etc...? If not, how can your technology know that a picture has an underage girl on it??

This is not a FantasyMan project, this is one of my and partners who have been innovating different things for years. I have never spoken to FM about this project.

Quote:

If Sniffy wrongly accuses someone as verified by your human eye low-tech factor, will "you" (in whatever capacity the "we" is in your statements) also stand as mediaries between the webmaster and the sponsor program who now has him labelled as a cheater and child pornographer?

We are not out to label anyone a CP profiteer, we however are lending our technology to the ASACP who have relationships with the FBI, we cannot store CP images unless FBI gives us a get out of jail card free, so for now we are rolling this area under the ASACP, and we are off doing other policing.

Quote:

When Sniffy flags someone for suspect content/keywords etc., is the human eye involved BEFORE reporting it to the sponsor, or does Sniffy do this automatically on "his" own?

We will report the findings to the sponsor in a means they ask, and it will be up to them to validate the areas of concern.

If we deep scrub affiliate programs we hope to be a 3rd party form of compliance that signifies the sponsors have done all they can to rid their programs of unethical webmasters.

Good questions Carrie :-)

Hi, this is Lee's partner in Sniffy....

What sniffy could do is detect if your images were used on another website and report that website to you. If a site resized the image, then that would make it difficult to detect since sniffy works by calculating a unique (hash) value for each image it finds. So if your pic is 1024x768 and someone resized it to 800x600, sniffy would consider it to be a different picture and wouldn't know how to relate the two together.

Most likely images will stay the same size (since most content providers produce web-friendly sizes) so detection of an image is very possible.

If the image were a sponsor banner, then sniffy could find the occcurences of that banner on websites since webmasters wouldn't change the size of the banner.

Think of sniffy as like a bloodhound. You have to give him the scent of what he is supposed to find. If you give sniffy the image to find, he can then search out and find it.

There are companies like Cyveillance who focus on protecting copyright/branding for a company.

Sniffy works similarly, but different, in that sniffy is geared towards validating sponsor's affiliate websites to ensure that the affiliates are not supporting CP related activity.


-dj

Sniffy sucks in all the images into a DB and a human reviews them. This software scans the websites for every word, and reads every word on the site...it does not do just a quick peek scan.

As our database recognizes the images and it has been validated as OK, a reviewer does not ever have to see it again, so over time as this database build out, it requires less human verification.

We hope the content providers will loan us their images so we can hash them, it would save us alot of wear and tear validating.

Anyhoo, we hope bringing Sniffy out is a good first step and gives our industry a tool to monitor who they do business with.

Everyone hates the thought of CP and hates the thought they get money from our industry, well we decided we could do something about this.

Now, we have to see if the industry supports our police dog.

This sounds interesting. I have killed a few affiliates who were pushing CP, and it would be nice to know if I had any more.

However, this may be the kind of thing that isn't a good investment for small guys like me. Unless the monthly fee was low, a lot of people would have to consider their margins.

Using keywords to find and block "bad" websites is what programs like Cybersitter and Netnanny are doing. These methods are not accurate and give false positive results.

For example, http://www.veganporn.com contains the keyword of porn, so it must be porn! But when you visit the site, it's anything but.

This example illustrates the problem with keyword-based searches.


Sniffy takes a different approach by downloading all of the images, and a team of human beings view the images and assign tags to the images that describe the image. Once the image has been tagged, it goes into the database and then can be used to find matches.

if pic A from a website was determined to be "adult content", and the same image was found on other, then sniffy can automaticallly detect that the image was "adult" since it was already verified, and also, can flag which website had the same occurrence of the image.

If a sponsor were to give sniffy their banner images to sniffy, sniffy would be able to report back which websites that were using the banner.

This is especially useful to sponsors who have affiliates that sigh up to their programs, but then stick their banner and link on many different websites and pages. While this may be legitimate action, since registering each website and page is too time consuming, sponsors may not know where the traffic is coming from. More importantly, they may not know if they are directly supporting a CP website because of their banner being on the page.

Sniffy not only verifies the list of URL's that a sponsor asks sniffy to scrub, but also casts a wider net in searching all adult related websites, and helping to find the other URL's that a sponsor's banner is on, that the sponsor didn't know about.

It's a lot of work to human eye validate the images, but it's necessary to ensure 100% accuracy. This is what the sniffy project is geared up to do, to provide a service that helps detect CP violations that go in parallel with ASACP's goal of removing CP from the internet.


-dj

Don't count yourself down and out as a small sponsor, at least your trying to manage your affiliates and check for unethical behavior.

We will be charging based on the size of the affililate program,

In the long run, sponsors who bring Sniffy on board will at least have some idea whats going on with their webmasters websites, and in the long time the ROI is lowered liability issues...

Kinda an ounce of prevention thing...is worth more than a pound of cure.

Webmasters who do unethical things don't need to be dragging our sponsors etc., into an investigation.

For sponsors to hire a full time person to police their affiliates would be expensive, so we hope our cyber police dog, gives them another alternative at a fraction of the cost.

This would seem a herculean task. Catagorizing hundreds of thousands of images by live people sounds very labor intensive and therefor expensive. As for how much we would pay, I suppose it really comes down to how much "bad content" it finds. Perhaps a bounty system as opposed to a monthly fee would be more appropriate.

You sir, are correct. It is a large task to validate the images. We believe through our custom software and the massing of a database that will help to auto-validate files that have already been found (therefore never have to validate the same image twice), will give us an inital high curve, but then goes downward over time. (we have already been collecting data, so the process is already in place).

As far as cost, we are working on the pricing, and most likely tieing it to number of websites a sponsor wishes us to monitor. The fee structure will most likely be a fixed monthly cost. The inital scan is the most resource intensive, and then as the month goes on, subsequent maintenance on new images are done, reducing the work load. In the end, it all averages out.

The problem with a bounty approach, is it would put us in a position where it's almost like blackmail... in that if we detect a site that belongs to a sponsor, then go to them and say, we found a site, you need to pay us $500 to find out, then it does put us in an uncomfortable position.

We would rather work on behalf of the sponsor in a very proactive way.

The pricing is reflective upon size of the web to cast, but in the end, we believe it will be reasonable.

In the end, it comes down to how much of a problem is this and how much a sponsor can afford to give them some level of protection of legal liability, as well as moral/ethical peace of mind, that they are not supporting CP activity.

Lee and I are very passionate about stopping CP, and have created this service as a way of combatting it, but the resources involved, do cost money. As sniffy gets closer to launching, we will be able to present the pricing structure, we are sure that it will be fair and worthwhile for the sponsors.

At this point, as the technology is finishing up, we are looking to see the interests and the problems of the sponsor community as well as our idea of the solution.

-dj

Any areas we overlooked sponsors? Any feedback would be appreciated.

No questions, nothing?

Hmmm...

This sounds like it would do ok at policing the free sites/TGPs/Linklists, but that is not where the majority of the problem is.. imo...

The majority of the CP etc is occurring in the member areas of paysites....

Does sniffy go in and flesh out the "undesirables" at that level ??

The problem with all automated scanning tecnologies is that they inevitably come up with false-positive matches.

But then again I see "human team" mentioned there. Lets hope this is a team of qualified LEGAL experts because anything else is just opinion re what is LEGAL and what is ILLEGAL.

LeeNoga, you better have a backup plan for any legal damage caused to any webmaster' reputations by false-positive matches.

Furthermore, I'd be careful about promoting anything being associated with a law enforcement agency (you use the term "police dog") unless you actually plan on hiring law-enforcement agents to inspect the images.

Maybe I'm being a bit of a wet blanket but you simply have to 100% think about the copy you write before you promote a product or service because in the long run anything you say or do here could be admissible in a legal action against you.

Good in concept Lee, but in terms of practical purposes I don't see how the logistics can function without a tremendous amount of false results with the enormous amount of content everyone has out there.

Again, you are promoting yourself as having relationships with law-enforcement agencies and the FBI.

Throwing around the term "Policing" in this manner is very irresponsible. Do you have authorization from either of these agencies, ASCAP or the FBI, to promote your new "policing" tool?

Obviously you have done your due diligence right?

Do you or do you not employ actual law enforcement agents or are you yourself an agent of law enforcement?

Think about your copy before you write it. Your service sounds like an admirable undertaking, and looks like a great start to bringing added self-policing to the industry..however, by positioning yourself as "policemen", you are opening up an entire pandoras box that neither your company nor this industry will welcome.

How do you figure that? Sniffy will report back based on a deep scrub based upon parameters the sponsor sets. The deep scrub being meta's, URL's, text etc....either you bust on a trigger or not, no false positives.

As for the image verification, thats the human element where the image will be judged based on specific criteria. If a sponsor wants to know who is pushing the 18 as younger....thats the sponsors judgement call.

Pretty black and white.

If we see questionable images and the site is not using known CP triggers, than its obvious this webmaster is not dealing in the lolita market.

Typically, if 18 yo presented to be younger is a webmaster goal they market to the lolita niche and the deep scrub wil pick up these indicators. NO grey area.

How can you say the following two lines in the same post?

We are laying that ground work now. ASACP is going to make use of our technology, special package "Sniffy Jr.". We would fall under the ASACP umbrella once they use it.

Quote:

Throwing around the term "Policing" in this manner is very irresponsible. Do you have authorization from either of these agencies, ASCAP or the FBI, to promote your new "policing" tool?

We just finished this software. Got to have a product before one can talk to the enforcement agencies. Our goal is to assist ASACP in finding the CP with Sniffy Jr., we with "Sniffy Sr."are looking to give Sponsors etc. a tool to peek under the hood of their affiliate programs.

Gary from Sex.com is allowing us to beta against his site. Very supportive of this product and its abilities etc.

Nothing like this has been developed, and its new and faces all kinds of scrutiny, but in the end we hope it makes a difference.

We believe once folks are more familiar with Sniffy, it will be looked as a wonderful step forward.

What do we have now? Nothing.

Its labor intensive, its mammoth with the database sizes, but what should I do sit on my ass and just bitch about CP and tell you how it makes me sick?

Nope. Doing something about it.

There are so many sponsors that have no idea what kind of traffic is coming into their sites.

As for the word police, perhaps thats a bad word, maybe I should have said, "Patrol" dog, "Watch Dog", anything but Police dog for we are not a law enforcement agency.

Quote:

Do you or do you not employ actual law enforcement agents or are you yourself an agent of law enforcement?

No, why would we do this?

Quote:

Think about your copy before you write it. Your service sounds like an admirable undertaking, and looks like a great start to bringing added self-policing to the industry..however, by positioning yourself as "policemen", you are opening up an entire pandoras box that neither your company nor this industry will welcome.

Not sure how you got stuck on us being the real "Police", prolly an encode and decode thing.

This is a tool that no doubt will attract the interests of various law enforcement agencies, its a "WATCHDOG", that better?

Ok, let me break it down.

Are we on the same page the entire site is scrubbed? Everything from meta's to text.....

OK.

If a site has underage pix, or legal pix marketed as underage, the typical webmaster will market to this lolita niche using keywords, alt-tags, text or something....or some kind of art graphic.

We are not here to say:

Mr. Sponsor, Joe webmaster has a young girls on his site based on the images we viewed. That would be our opinion and thats not our job with Sniffy.

However, we could say:

Mr. Sponsor we found one of your webmasters that use keywords and text to market underage girls. The images on the site could be underage or of legal age, but if they are legal they are being used to market underage content.

So with us there is no grey area.

So the "human" viewing the cp images will be a law enforcement officer?

Get whatever terminology you plan to use cleared with a lawyer.

I still am left confused how you can do this without yourself illegally accessing cp? Which is why I cant understand how youd be able to view and scan sites for cp unless youve got a relationship with a law enforc. agency.

Please explain. I just woke up and im a little groggy. I apologize if Im not getting it.

If all it takes is to change 1 pixel, how long before scammers realise (i'm sure they read the boards)?

Mr. Popup.

I agree with your posts and the concerns you have expressed.

There is a very serious legal issue facing sniffy in the area where human reviewers are viewing CP material. While we hope that 99% of what sniffy finds is clean, it's that 1% or so, that could get us potentially in trouble.

We are addressing this legal issue by working with ASACP who has the FBI relationship established. Only people who have been sanctioned by the FBI as "CP informants" would be used to do the reviewing (just like the way it works today). These people would not only be sanctioned, but also have the stomach to view these kinds of images.

As Lee wrote, we are working with ASACP in this service. While sniffy is a commercial venture, it does share the same goals as ASACP and with that synergy, we will have our legal side covered.

It would be an ironic shame to have such a service that has the righteous of cause, be shutdown and jailed for CP violations. Neither Lee, our team, or I wish any such outcome, so your concerns are taken seriously and to heart.

-dj

An excellent point! You are correct, that even a 1 pixel change, or even a change in the jpeg header for comments, can cause the image to be considered a different image.

If a person altered an image in the hopes of hiding from sniffy, they still have to make that image available via the web. If it's on the web, then sniffy's aim is to retrieve it.

So this new image with just the 1 pixel altered won't get automatically scrubbed against the database, but when the human reviewer sees it, they won't realize that it is the same pic in the db that has just 1 pixel changed, but instead, will judge the picture based on a set of guidelines that determines if the image is "adult content", or CP, etc.

Since sniffy spiders from multiple locations (differnet ISP and ip numbers), it makes her harder for a webmaster to try to foil sniffy.

There will certainly be many challenges that we and sniffy face, but if you agree that the issue of CP is a problem, then we are offering a possible solution. Whether it is successful or become the defacto standard, is something that time will tell.

Lee and I feel we have a solution, and believe that we can do something proactive about this issue.

Your comments and feedback are very welcomed. You certainly help us to flush out our vision in a public setting so that our efforts become more communal, where we all are sharing and moving in the same direction of CP removal.

-dj

dj, Lee....

I think the question I asked got lost in the mix so tospeak, so I will ask it again...

I realize that you are targeting this product at the sponsors... But as, imo, most of the real CP problem is occurring at the paysite level, how is this product "truly" going to help with the CP problem....

Glad I read the board before posting back.

Thanks DJ for jumping in there and answering the questions quickly.

Its a huge undertaking, and it has taken months to write the code and months to build this huge database out, we are up to 5 million NON CP images...

We hope as Sniffy is embraced, content providers will loan us their content so we can hash it and load our database, it would cut down on human viewing etc.

We feel this is a step in the right direction. Affiliate software years ago was not ready to handle webmasters with 100+ sites, and one problem we were told of is the sponsor has no idea where their banners are flown from.

We have cross referencing abilities too, and can tell sponsors where their banners are located....

We have a long road ahead of us, but we are committed to taking that walk.

We have already overcome alot of anticipated problems with poison pages & member areas.

If the sponsor gives us the members passwords, we can sniff inside. AVS sites or notorious for being problematic.

Yes, its a trust issue for sponsors to give us their list of known URL's/passwords.

But, since my dealings with webmasters go back to 1995, those who are sponsors know me and know what I stand for which should desensitize this area of "trust".

I am proactive to a fault when it comes to ethics in this business.

Great questions and feedback, BTW. :-))))

I am sure Dj is gonna respond too to this one.

Sniffy can be targetted at Sponsors to scrub their affiliates, it can be used by Search Engines, it can be used on Link sites like "PersianKitty" etc. Sniffy can be used on any project that involves a working relationship with multiple websites.

CP member sites are pay and secretive, Yes.

However there are marketing pages out there to help people find their way to these locations, we will locate those paysites via the marketing trails leading to them...

We hash every banner and every graphic on every adult website we crawl. Finding the sales pages to get business to the CP paysite is something Sniffy can do. Remember, we are sniffying for triggers in text, keywords etc....those that are heavy in the words [free sites] no doubt will point us to the CP paysites.

Granted we will turn these over to ASACP unless the umbrella extends over Sniffy too, they will be using Sniffy Jr [ASACP will].

Hope this answers your question :-) Great question!

This sounds great except do you plan to pay me Mr webmaster the bandwith you are using by sucking down all my pictures only to prove I own the rights to use all of them?

Also can't they just do a robots not allowed or block the spider by ip/stamp with htaccess?

Some technical perspective here:

A good idea, but you don't have the technology.

Your tech guy said you are hashing (md5ing) the images. A simple resize kills your review and creates a duplicate entry.

You know what else does? Adding a url. Cropping. Color correction. Slight recompression.

You are going to have a database of millions and millions of pics, a large percentage of which will be dupes.

Invest in some real tech (not md5 hashing) and try it with that. By real tech I mean that there's some very high level image recognition tech that could probably do it - but it's an investment.

I don't see this being worth your time to be quite honest.

Sorry for the downer, just MO.

On the first question of the bandwidth issue. You are correct that Sniffy will be downloading images off of your website and that will incur some kind of bandwidth charge, but it is no different than a surfer that has found your site, went through your pages, viewed all of your pic of the day, or guest pics, then decided they didn't want to buy anything you had to sell.

You don't and can't charge these people for "wasting" your bandwidth, it's all apart of doing business.

Sniffy will be responsible and a good visitor by not overburdening your servers.. it acts just like any surfer and will not cause any more harm than a normal surfer.


On your other point about trying to block. Yes, a webmaster could try to block us, and of course that would just raise our suspicions as why someone would want to hide.

Since sniffy spiders from several ISP and IP's, it would be hard to block it. Sniffy is smart enough to get around referer and IP type blocks, in order to do its job.

I would hope that only a small handful of affiliate webmasters would be concerned about sniffy's actions, since they would be initiated on behalf of an affiliates sponsor, to ensure compliance with THEIR policy of no CP sites.

Sniffy is not meant to be a big brother type operative. We would hope that fellow webmasters would be supportive of any measures that could help reduce CP.

-dj

As mentioned in a previous reply, you are correct, any single bit of manipulation of an image (other than renaming the file) will cause a different hash value.

Yes, it does mean there will be a database of millions of pics (there already is). The data will never be dupes...only the unique hash will exist. It is possible that a 1024x768 picture from a content provider, gets resized and manipulated 5 dozen ways, resulting in 5 dozen versions of the file, but this means that those 5 dozen versions of the file exist on the internet, therefore are viewable in a web browser.

So with sniffy's activities of being a good surfer, and not putting a load burden on a website, it will find these images, that to the human eye would appear the same, but at the bit level, are uniquely different.

Yes, it does mean lots of images to validate, but that's really our problem to deal with.

There is alot of technology that sits behind the suite of applications that process the data, and it's combined with alot of elbow grease to make everything work.

Pixel analysis in looking at flesh tones is probably what you are talking about using real technology..and i know there are companies out there pursuing that path.

In tests with tone detection as well as the keyword type blocking, there is still room for error. When a person validates an image as being "adult content" or "cp", given strict guidelines, it is 100% accurate. There are no false-positives that could exist, as it does today in anti-spam software, pixel flesh tone analysis, and keyword searching.

We could most certainly incorporate flesh tone analysis programs into our suite of tools, we certainly aren't beholden to our own applications, and this idea has already started to be explored. Anything that speeds up the task of providing 100% accuracy is in our best interests to explore.

thanks for your thoughts, they are appreciated in this discussion about trying to tackle the problem of CP on the web.

-dj

We are not about hassling webmasters and intellectual property right issues.

If you want to use 18 year olds as content, more power to you. But if you market lolita all over your site and hope your content is viewed as under 18, more power to me :-)

Those that want to trick the surfer into thinking they have underage content [knowing they have model releases to show otherwise], is of some concern to sponsors.

I did not say flesh tone analysis - that's a different thing altogether and probably wouldn't be useful to you. It's about determining whether an image is porn, not comparing two images.

High tech and low tech? More like low tech and really low tech. This is something your average TGP scripter could write. Seriously - it's a spider that md5s pics and puts the hash in a db with a little extra data that the army of reviewers inserts telling you whether this particular variation of a pic is CP or not.

Weak.

Hmm.. To be devil's advocate here.. This is starting to sound very Orwellian to say the least...

Because someone blocks it, they must have something to hide ?? Thats the tactics the cops use to get in to search your car... "What do you have to hide?".. Well in this case, not a damn thing, but it is my right, as payer of my server expenses to control my traffic anyway I choose.

It may not mean to be a "big brother", but quite frankly, I have the right to run my sites the way I want to... If the only thing that sniffy is doing is checking for cp, you might be able to sell that... But, that is not the only thing it is doing... It is actually creating, basically, a roadmap of my sites, my content and my templates....

I work with many more than one sponsor. What you are doing here is "potentially" giving a sponsor the ability to check out everything I am doing, what sponsors I am using, who I am selling/sharing my traffic with. What is to stop you from selling that info to the sponsor... Then if one sponsor starts playing war with another sponsor, what is to stop them from cutting me off, or threatening me.. "hey, you better lose sponsor so and so if you want to use us as a sponsor"... What is to stop you from passing all my domain and site info to a particular sponsor...

The sponsor might have a right to access a gallery that I am sending them traffic from in order to check on me, but you are compiling an entire database of all my domains/sites/galleries. What is to stop you from misusing this info. Your good name ??

Further... You are compiling perhaps an entire catalog of all adult web content... What security will you have in place to keep hackers out of the database? If all the images are stored in one place, it makes it a nice target for those that want a shitload of images on the cheap and easy.

Using multiple ips to run the bots from is no problem, but using faked referral strings, etc to subterfuge the .htaccess is. These are standard hacker techniques often used by unscrupulous webmasters to hotlink and steal content. If you are using the same techniques then you are the same as them and I feel that I would have the right to block you for the same reasons. Just because your cause is "righteous" does not allow you to use the same tactics. The ends CANNOT be used to justify the means.