3
$\begingroup$

These guys here: https://www.patreon.com/AiAngel are saying that they've created a AI who can chat and stream. As the so-called administrator "Rogue" said:

enter image description here

this chat/streamer bot are no fake. Also, there's more about the dynamics of this chat/streamer bot on youtube:

https://www.youtube.com/watch?v=WyFwjHQhlgo&t=463s

https://www.youtube.com/watch?v=GtvivssqLhE

Considering that videos I realy think that this bot is totaly fake. I mean, I think that even the most advanced AI bot do not get even close to a real conversation like this one.

Now, of course that you can say that this is a artistic project or something, but the people behind all of this are on Patreon, and the people who are paying to these guys possibly are getting totally fooled, which is a serious thing when we're talking about real money.

So, is AIAngel a real bot? (With this question I'm spreading this possible fake to community)

$\endgroup$
  • $\begingroup$ Thanks for asking this question and welcome to SE:AI! (I've adjusted the header for searchability--hopefully anyone googling for more info on this supposed bot with find your question!) $\endgroup$ – DukeZhou Apr 29 at 18:17
  • $\begingroup$ Neither the quote you give nor the videos claim that AIAngel is using any AI technique. At worst, it seems that it is being deliberately left open in order to trick the unwary (based solely on the name). Could you link to any claim that AIAngel is any kind of AI is by the author? As far as I can see, it is just a modern YouTube version of Max Headroom - see. en.wikipedia.org/wiki/Max_Headroom_(character) $\endgroup$ – Neil Slater May 6 at 18:29
3
$\begingroup$

To answer the question if AIangel is fake or real we have to describe first what the idea is behind a human-chatbot-conversation. A well known chatbot is Eliza (created in 1966, easy for reverse engineering).[1] The human can do the conversation with Eliza in two ways. The first one is to assume that Eliza is a computer program and the human has to show that the bot doesn't pass the turing test. For example, we can ask Eliza “what is the weather today?” and if the software doesn't know the answer, then it's fake. The other interaction option is to estimate that Eliza is a real human and try to do what the chatbot wants.

The same interaction situation is available in case of AIangel. Asking if the shown avatar is a real human or not is only one possible way to play the game. The more elaborated experiment is to assume that it's a real human and try to investigate what the needs of AIangel are. We can ignore the turing test and go directly into the conversation. Don't ask what the chatbot can do for you, ask what you can do for the bot.

[1] Pereira, Maria João, et al. "Chatbots' Greetings to Human-Computer Communication." arXiv preprint arXiv:1609.06479 (2016). pdf

$\endgroup$
1
$\begingroup$

Without being able to interact with the bot, a Turing test is impossible.

  • Based on the manner in which this supposed bot is presented, in videos as opposed to an interactive medium where users can interact with the bot, the only reasonable conclusion is hoax.

This assessment is supported by the overly sexualized rendering of the bot combined with requests for donations.

In order for this project to be considered legitimate, the creators would have to be more transparent about the methods and allow the general public, or at least reliable experts, to interact with the bot.

(Compare to IBM's Tay and Zo.)


Turing Tests & Pornbots

The nature of Turing tests is that they are subjective. A bot that an adult human could easily recognize as non-human, a child might perceive as human--the child does not have the requisite knowledge to form a strategy to expose the bot as an automaton.

There is an idea that "pornbots" have been passing the Turing Test for many years now, predicated on the hormonal imperatives of those interacting with the bots, which greatly inhibits their judgement. (The purpose of pornbots is to get consumers to spend money. Their continued existence suggests they do have utility in this regard, and that utility can be seen as a confirmation of the bots passing the text in a specific context.)


Automation Hoaxes

One of the most famous machine intelligence hoaxes was The Turk, presented as a Chess playing automaton. The supposed machine astounded late 18th century audiences with its skill, but it was later revealed to be a human in a box.

$\endgroup$
  • $\begingroup$ Whilst I agree that the chatbot is not driven by AI (at least as far as perception and conversation, where it appears to be just some kind of rigged avatar), as far as I can see the author makes no such claim for it, at least not from anything linked by the OP. Which makes it difficult to declare a hoax here. $\endgroup$ – Neil Slater May 6 at 18:58
0
$\begingroup$

No, AiAngel is not a bot. It's Rouge's software that changes his voice along with facial recognition software which tracks his movement and copies it to the avatar.

That being said, he has created a very entertaining channel and line of work for himself. By looking at the videos, you can see that he is a true genius at work. Single handedly puts on all hats for that project, from hardware infrastructure development, to software design to graphic art to acting and video editing. Truely one of the great minds of this era.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.