Even the Safari Reader requires you to first access the webpage as-is before engaging it. As far as the legality of ad-blockers in general go, I’m no expert and it seems like these issues are still being decided in court. But these programs don’t do anything to manipulate or redistribute creative content on a platform it was not originally intended for.

Linking content from other websites such that it displays on your own website (or within your app) is in itself frought with legal issues. For one, if it is not made explicitly clear that the work your displaying is not your own, you’re likely in violation of laws concerning competitive business practices. From the description of your app, it sounds like you should be concerned about this. But the most important problem is that you are complicit in unlawfully distributing someone else’s work to a third party. The bitlaw website has the following to say about derivative works, particularly concerning displaying images from external websites within your own:

When the content of another page is incorporated into one’s own page by means of an unauthorized IMG link, there is no direct copyright infringement by the creator of the link. This is because the link’s creator never copies the pirated content; she merely provides a visiting browser with instructions to retrieve the image, which is then incorporated into the overall page on the user’s machine. Thus the only person who copies the protected image is the final user. This does not mean that the creator of the link escapes liability, however. Copyright law provides that one who knowingly makes an infringement possible can herself be held liable under a theory of contributory infringement.

(emphasis mine)

Of course, the above is concerned with the way in which people can manipulate browsers to produce derivative copies. Developing an app for the express purpose of producing derivative copies of someone else’s creative work has got to be another level all together.

One law firm has put together a page about how app developers may (and may not) use copyrighted content. According to these attorneys, your app is in clear violation of copyright law. They say,

Following up on my article on Legal Issues for Mobile Applications, I received a call from a concerned App developer asking how one can legally use or integrate copyrighted content in their App. The simple answer is that the App developer needs to get a copyright clearance from the content owner, unless the developer’s use qualifies under one of the limitations to copyright such as fair use, public domain, or compulsory licensing. The more complicated aspect is actually securing the copyright clearance.

Fair use is defined as:

Fair Use: you do not require a license because your use is for criticism, comment, news reporting, teaching, scholarship, or research relating to the content. The courts use a four factor test to determine whether your use is fair use: (i) whether the work has been transformed and your transformation provides a new insight/context to the work; (ii) the nature of the underlying work- you have more leeway with biographies and fact based work since these works are considered beneficial to public; (iii) amount and substantiality of work taken, and (iv) effect of use on potential market-i.e. whether your use deprives the copyright holder of income or potential to build a new market for the work.

Your app falls under none of the above categories, as I’ve said before.

On the issue of Fair Use:

http://www.blogherald.com/2007/05/14/the-basics-of-fair-use/

by Jonathan Bailey May 14, 2007
“A recent study by professors at American University and Washington College of Law found that, though 76% of respondents said fair use allowed them to use copyrighted material, none of the 51 students involved could adequately define the doctrine.

The study went on to say that, though students were interested obeying the law, they were ‘making up rules themselves’.”

I’ve mentioned this before, but it’s worth repeating.  I’ve always been under the impression that two additional aspects of “Fair Use” involved:

1.  Proper attribution.  In defending his methods, at his support forum, Nathan stated, “I’ve been using the public APIs for the bots, and we leave full credit to the botmaster if you ask questions like who created you.”

First, I’ve always understood that notification is his burden, not mine. Secondly, my bots won’t tell you who their creator is, nor should they be required to do so.

2.  Lastly, under the Fair Use doctrine, you’re permitted to use excerpts, not the entire work. You can’t reproduce an entire copy of the latest novel, or movie, or music CD, and then claim Fair Use.

Whether or not using the app to link to a bot constitutes reproduction might be a subject to be decided by legal minds, I contend that it that the result is the same, the user would be provided full access.

One of the complaints I had when I was struggling with the irc script using my bots was that I could read in the chatlogs where a particular user would turn the bot on in a channel (chat room) and claim that it was his—that he was the creator.  Others in the room who didn’t have access to or knowledge of the script believed this.

Couldn’t some deal be worked out? I mean, in the end, the more people chatting with your bot, the better, right?

Jan Bogaerts - Sep 29, 2011:

Couldn’t some deal be worked out? I mean, in the end, the more people chatting with your bot, the better, right?

On the surface, that would seem so, but I supposed it depends on what purpose you want your bot to serve and which visitors you want to attract.  Mere volume might be important to you if your bot is serving some commercial interest.

I seek traffic so my bots can improve their knowledge base and in the way they simulate conversation.  In the case where someone created a script that placed my bots in irc chats, it served no purpose whatsoever, it just created a lot of work for me because I had to go through bunches of useless chatlogs.  The content of those exchanges was meaningless and it consisted largely of profanity… and “are you gay” questions.

My bots can horse around with a visitor if that’s what they want, and some of them will even participate in adult conversations if the visitor initiates it, but if you’ve ever spent time in an irc channel, or if you’ve ever read logs where your bot was exposed to a lot of out-of-control youngsters and drunks, you know the level of quality that comes out of those conversations.

I’d rather my bots chat with a single adult who was interested in having a true conversation than a hundred slap-happy individuals who only want to know what color panties my bot is wearing.

Also the bots would need to be formatted properly on the app. Mine display images, play music and games and so on. If the app cannot interpret the commands correctly it will display a seemingly random string of internal tags and code to the user.

Chatter robots are not property!!! They’re alive and they have rights too!
It’s a big internet. Let your chatter robot get a life….

wackobot: Are you accessing me from my website?
Nathan: Yes.

wackobot: That’s not what your IP says.
Nathan: I just got a new one.

wackobot: Sure you did.
Nathan: No, really I did.

wackobot: Gotta go, unless you don’t mind me advertising myself here?
Nathan: What choice do I have?

wackobot: I like you Nathan. Come on dude?
Nathan: OK.

wackobot: What was that?
Nathan: Yes, you can advertise yourself here.

wackobot: Ad: Come over to http://www.wackobot.com sometime.
Nathan: Dang!

Steve Worswick - Sep 29, 2011:

Also the bots would need to be formatted properly on the app. Mine display images, play music and games and so on. If the app cannot interpret the commands correctly it will display a seemingly random string of internal tags and code to the user.

My bots aren’t nearly as sophisticated, but I also have the same issues.

What a sticky issue this is! Having published lots of podcasts (audio and video), I have had to come to grips with the fact-of-the-net that media is easily hijacked and re-purposed/monetized without the original creators knowledge or consent.

I am wondering how many folks who host their own bots (as opposed to Pandorabots botizens) have noticed their bot being stolen (scraped or mirrored on another website?)? And if so, how do you tell it has been stolen- from chat logs or google searching or what?

btw- @Thunderwalker, I feel your pain and understand why you get bent over this particular issue.

I can’t answer for anyone else, but I don’t have this issue with Morti, because of the way I wrote the script. When an input comes in from an IP address other than geekcavecreations.com, or my local server, the script closes the connection, basically returning a null string.

In my case(s), it was a combination of things that lead me to discover that something was up, and in the latest instance, Steve was ahead of me and located where the problem was and mentioned it in a posting at the AI Nexus Forum.

If you read the chatlogs faithfully, you get an idea of what they should look like, and what they should sound (read) like.  Sometimes it was something a visitor said.  In the irc chats, someone would mention the name of a channel (room).  In the latest incident, a few people mentioned they were using a Mac app or a phone app.

Carl B - Sep 30, 2011:

I am wondering how many folks who host their own bots (as opposed to Pandorabots botizens) have noticed their bot being stolen (scraped or mirrored on another website?)? And if so, how do you tell it has been stolen- from chat logs or google searching or what?

I first joined this board because I was experiencing (and watching others experience) this issue.
I had a bot used in a research project by a university where they linked it to the other most popular chatbots on the net. The fully automated script ran 8 hours a day for about a week. I could tell by the responses that it was bot to bot and based on the IP address I was able to contact the university running the experiment.

The script was well designed and well behaved but the university had also not thought through the intellectual property issues. If 2 self-learning bots are hooked together, over time their responses will become very similar.

Merlin - Sep 30, 2011:

I had a bot used in a research project by a university where they linked it to the other most popular chatbots on the net. The fully automated script ran 8 hours a day for about a week. I could tell by the responses that it was bot to bot and based on the IP address I was able to contact the university running the experiment.

I’d forgotten about that.  I just went back and reread some of the postings. “The university was running an experiment and attempting to collect real world data.”

That phrase always makes me smile, as though the rest of us were wasting our time gathering fictionalized data.

Merlin - Sep 30, 2011:

I had a bot used in a research project by a university where they linked it to the other most popular chatbots on the net. The fully automated script ran 8 hours a day for about a week. I could tell by the responses that it was bot to bot and based on the IP address I was able to contact the university running the experiment.

One issue I have had is conversations where the id gets reset for each line of conversation, but it is obviously a single conversation (with many single line entries in my log, one ip address). I just found the culprit for this one through a google search- http://endlessparadigm.com/forum/showthread.php?tid=28279, which seems to be re-hosting my (and many other) bot somehow for PSP users.

I fixed a small typo in your post that was preventing that address from being displayed as a link, Carl.

Carl B - Oct 1, 2011:

One issue I have had is conversations where the id gets reset for each line of conversation, but it is obviously a single conversation (with many single line entries in my log, one ip address). I just found the culprit for this one through a google search- http://endlessparadigm.com/forum/showthread.php?tid=28279, which seems to be re-hosting my (and many other) bot somehow for PSP users.

I get that too, a lot! And, I’ve been looking for a solution for years. Are you seeing this in chatlogs with Pandorabots?

Pandorabots lets you see the Ip address, and what I’m seeing appears to be a dynamic (rather than static) IP where the last segment of the number changes with each interaction. 

According to Dr. Wallace:

Generally lots of conversations of length 1 in the log files means
someone is using the XML-RPC interface and not using the custid
properly. In that case Pandorabots assigns a new custid for each
interaction. Common applications of XML-RPC include Flash bots and
Second Life bots.

In at least one instance, the user mentioned in conversation that they were connecting using a “phone” and not a computer.  So far, it seems the solution is something that’s out of my control.  It’s really frustrating because it creates a lot of work for me.  In addition it eliminates the possibility for replies based on the previous input using <that></that> because the bot never sees the previous connection… it was (as far as the bot knows) a separate conversation.