Forum Moderators: open
BT should face prosecution for its "illegal" trials of a controversial ad-serving technology, a leading computer security researcher has said.Dr Richard Clayton at the University of Cambridge made his comments after reviewing a leaked BT internal report.
The document reveals details of a 2006 BT trial with the Phorm system, which matches adverts to users' web habits.
"It's against the law of the land," he told BBC News. "We must now expect to see a prosecution."
AD Trial Brings Calls For Prosecution [news.bbc.co.uk]
Earlier stories
[webmasterworld.com...]
[webmasterworld.com...]
[webmasterworld.com...]
[webmasterworld.com...]
[webmasterworld.com...]
Anyone who runs advertising or is in an advertising neetwork, should watch this one like a hawk. It is the life blood of many small websites. If true, I think what British Telecom did, was nothing short of stealing from small websites.
Not to mention the big boys (like Google) that provide the Ads for smaller sites. I'd imagine that their Lawyers are already preparing the case.
[webmasterworld.com...]
(and mentioned some leaked documents from Wikileaks about the BT tests on page 4)
What's worrisome is that some smaller ISPs might already be using the deep packet inspection technology but not disclosing it, thereby potentially nibbling away at our publishing revenues.
If there's a script or something that we can add to detect the technology, that would be great (however given that DPI allows ISPs to modify the HTML of a page, conceivably those ISPs could remove the script!).
If this isn't theft then the definition of theft needs to be amended, by statutory law, to include this practice.
[edited by: Webwork at 2:57 pm (utc) on June 6, 2008]
We only serve ads to the websites we partner with. In order to participate, websites have to insert a tag into their own page.
But I'd still rather it died a death. Give them an inch.....
During the trials adverts were stripped out of web pages served up to BT customers and replaced with more targeted ads, if available
I hope that means they were stripping out and replacing ads from sites that are in Phorm's network, but I am suspicious to say the least.
rj87uk, I told you so!
If there's a script or something that we can add to detect the technology, that would be great (however given that DPI allows ISPs to modify the HTML of a page, conceivably those ISPs could remove the script!).
This should be possible, might be as easy as generating a hash on the server side to represent the final HTML page and then creating another hash of the page with a javascript function once the page lands in the users browser, then sending it back to the server for verification in the background. The server would notify the webmaster if the hash differed and the webmaster can decide to take action.
Does the above statement mean that much of the damage this system could cause has been mitigated? - it still sucks though -and of course policies can change overnight if they are not making enough money.
We only serve ads to the websites we partner with. In order to participate, websites have to insert a tag into their own page.
Ok so now I am slightly confused, Does this mean that only websites that are using the "advertising" program will have ads uploaded onto a site that has choose to run this program?
Im confused now.
According to the articles and the document though it appears they replaced ads during a trial run but there is no indication that is what they are doing now.
A BT customer who has opted-in to the 'Webwise' system will be monitored everywhere he goes on the internet. Every standard page that he reads and writes to (https excluded) will be scanned and a profile of that user created.
If he then visits one of the partner sites (ft.com, ivillage...) he will start to see ads based on his profile.
The initial forecasts where BT could earn £85 million were based on the system being opt-out only. The ICO smacked them down over that so they have had to re-write the system to be opt-in and the revenue forecasts are now expected to be well below what was initially expected.
As for the trials in 2006 and 2007. Well we have had a lot of spin and lies and only now the truth is being revealed and hopefully the ICO, Home Office or the EU if need be will now do the right thing and take action.
As for the future - they can make the system opt-in, DPA, RIPA and PECR friendly as much as the can from an ISP customer point of view but there is still the issue of us webmasters having our copyrighted website information read and processed.
As I said in this thread:
[webmasterworld.com...]
All Your Content is Fair Game
If a visitor to your site is Phormed all your content is going to be read and processed whether you want it to or not. This will include parts of the site protected by (non https) .htaccess or php or messageboards where registration is required. The only way to block Phorm is to block Googlebot.
Phorm is Going to Direct Your Customers to a Competitor
A Phormed visitor browses your site. This means that your site is helping to make a profile for that visitor. If your site is about exotic holidays then Phorm is going to start showing that visitor ads for your competitors exotic holiday websites.
Google: Uses keywords from your page content to bring visitors to your web site.
Phorm: Uses keywords from your page content to draw visitors to someone else's web site.
So it is an "optin" type system
The technical spec says opt out, for which you need to acquire, and retain a cookie on your machine. Lose the cookie - you're opted in again.
I imagine customers will be 'encouraged' to 'opt in' via updated terms and conditions at their ISPs. So, the only other opt out will be by opting out of the ISP altogether, which I will do if my ISP goes with Phorm.
I will also inform my ISP as to why I decided to leave, and I'll post about my decision on relevant websites. I would encourage others to make their feelings about this system clear. However, in my business and professional dealings with the major UK ISPs, I've found their general approach to customers to be (almost without exception) dis-satisfactory in the extreme.
Any smaller ISPs reading - advertise yourselves as 'phormless' and you could get a good little niche going ;)
The only way to block Phorm is to block Googlebot
This isn't true. The researcher named in the article posted above has been given unprecedented access to their system, and they have given him permission to post details of the system. You can detect and block phorm with varying degrees of efficacy as both site owner and internet user.
"If I do not opt-in will I still be profiled"
Not opting in will only stop the user being served OIX ads. His profile will STILL be made - he will still be monitored.
I believe they may be working on a fix for that and it could be the reason why trial #3 is beinc constantly delayed.
---
Now back to the member area/piggybacking the traffic. This is a serious concern I have as one of my sites has 'finacial information' where subscribers can analyse data, talk in the non public domain message board etc.
Some phorm huggers say the onus is now on me to convert to full https but why should I do that? This is a problem I never had before.
My security systems has worked fine and keeps out freeloaders and such but there is no way I can keep out a Phormed user who will be leaking my and my customers private information.
I kid you not that it was suggest that we have to block googlebot in robots.txt if we want to stop Phorm from profiling our sites. The response from BT/Phorm was that any site which lets google in is deemed to be in the public domain and that if you let google in Phorm have every right to read that same data.
Of course that is totally unfair as I choose to let google crawl mysite in order that I can have the benefit of increased traffic. Allowing Phorm to analyse my pages will give me no benefit and may actually give benefit to a rival site who has signed up to OIX.
And of course - I don't let googlebot anywhere near the members area, or the parts of the messageboard which require registration. Yet Phorm can waltz in and analyse the pages.
" You can detect and block phorm with varying degrees of efficacy as both site owner and internet user"
As a site owner I think it is impossible to block it as the packet inspection is done at the ISP level - of course you can block those IPS's who allow it via IP address, but in the UK you are blocking Virgin and BT which is a large % of the UK market - not really sensible.
if you let google in Phorm have every right to read that same data
But this is an attempt at justification, not technical advice. Perhaps a quick breakdown of the tech would be useful.
Here's where it gets interesting. The cookie is set on your domain name (apparently named 'webwise', the name they given the system in an apparent attempt to make it seem like a security feature).
If (as a user) you block the cookie, they temporarily stop the phorm redirect (otherwise it would go on forever).
The reliance on cookies set on third party domains is their main Achilles heel. Users can block them, and site owners can detect them. I've heard that blocking the cookies as a user may result in performance problems with your internet connection. As a site owner, you can choose what to show to visitors carrying a phorm/webwise cookie. Of course, you shouldn't need to bother, but that's a different discussion.
Further, from what I've read the system seems like it could be vulnerable to hacking, due to the implementation. I imagine it will also be an attractive target. I'm a bit dubious about some of the ways they seem to be handling redirects, and detecting problems.
Added: just saw your post, driller41, but see above.
[edited by: Receptional_Andy at 8:19 pm (utc) on June 6, 2008]
There will be guest speakers at the event and later in the afternoon a case file is being presented to the police.
More information here:
[nodpi.org...]