Re: I thought this had been done and deployed ...
Ah... I look forward to a blissfully sparse Register comment section.
As if the internet isn’t already a complicated cesspool full of trolls, AI engineers have gone one step further to build a machine learning model that can generate fake comments for news articles. The eyebrow-raising creation, known as DeepCom, was developed by a group of engineers at Beihang University and Microsoft, China. “ …
Paul Kinsler,
But I wonder if you could use it as a filter; i.e. test user-generated posts against outputs from the bot, and if the match is too good ask the user to improved their post ...
Fuck off! Where else will I get to post my "jokes" but the El Reg forums?
Then I'd be forced to talk to real people. The consequences are just too horrible to contemplate!
For example, oppressive regimes could implement such a model to automatically dump a load of fake drivel to drive propaganda. The comments could also kickstart toxic arguments between bots and humans to sow discord and misinformation. Perhaps miscreants might even use it as a way to advertise products or post spam.
Well, that's half the commentards here out of a job...
I thought that the likes of Google and Facebook had bots that did this yonks ago?
Perhaps this is Microsoft catching up (yet again)?
All those 'I am not a robot' capcha's have already been irrelevant for some considerable time.
Then there was an article about how you can't be sure that you are playing a human being at Fortnite from a few days ago.
We are doomed I tell ye, doomed!
Could the same effect not be achieved by selecting a comment at random from a small pool of options?
e.g.
"This raises interesting ideas."
"Surely nobody believes this."
"You couldn't make this stuff up. Oh wait somebody did !!!"
"At last !!! Somebody gets it !!!!"
"I blame {Facebook|Sunspots|Twitter|Aliens|$POLITICIAN|$EVENT}"
"How does any of this help the environment?"
"Where's the IT angle?"
The number of accompanying punctuation marks, errors of spelling or grammar, quotes from the article and inappropriate capitalisation could also be randomly generated.
Actually, this could work to our (ordinary mortals) advantage. We already have bots to generate fake news, and we now have bots to generate fake comments.
Why not just let them talk to each other and we can get on with life/listening to music/useful stuff, probably away from the keyboard?
There is a sci-fi series called Altered Carbon, where there are AI run hotels that are empty because no-one really likes them because of their weird quirks and people have gotten over the gimmick. They sit there empty for decades just maintaining themselves and awaiting customers (who almost never come).
Have you been on the internet lately?
Well exactly. I was thinking that in my first days online in the 90s someone posting either:
The USA were rather late to WWII (7-ish words)
You'd all be speaking German if it wasn't for us (10 words)
Was enough to derail many online conversations. And a well placed "Fake News" or "Bollocks to Brexit" is pretty effective nowadays.
"Look, I came here for an argument!"
"Oh sorry. This is abuse. You want next door."
If you've seen "bored panda", judt about every photo has a pointless comment by some random user that must be done this way.
https://www.boredpanda.com/text-design-fails/, for example.
And take a look at the "reviews" on Source forge. Typical - "nice app!"
I am old fashioned, but I expect a review (or a comment) to impart some information I can consider making use of in forming a judgement of my own. However, as far back as 1942 Erich Fromm wrote "We are proud [...] that we are free to express our thoughts and feelings, and we take it for granted that this freedom almost automatically guarantees our individuality. The right to express our thoughts, however, means something only if we are able to have thoughts of our own."
This is becoming difficult as the information we can gain access to is increasingly both homogenised and "personalised" by crude automated filters under the control of faceless behemoth business. So this comment generator is just another small step for mankind. We'll all finish up like the protagonists of E. M. Forster's "The Machine Stops" (1928).
Like, build a few hundred thousand commentor instances and set them randomly reading stuff on the internet and talking and arguing about things and voting on which arguments are the best at changing minds and maybe breed those agents more and run that for a lot of time on a really powerful supercomputer see what happens
surely not in China?!
...
but I blame Trump. Yeah, seriously, I blame Trump who has proven that the leader of a world empire, this shining example of human behaviour, can say anything, in public, any lie - and the world listens and emulates, including all the little hitlers and tinpot dictators (pardon me, democratically elected leaders in world's best democracies). So, what's so outrageous now about academics in China proudly announcing to the world that, on top of fake news, fake comments are great? Great, GREAT MOVE I WISH THEM ALL THE BEST.
"The paper didn’t mention any potential malicious applications of the technology"
Yes it did. Literally every possible use for it they mentioned is malicious - it's all about posing fake comments to trick real people into thinking an article or site is much more active than it really is. What the paper didn't mention is any possible legitimate applications. Because there aren't any.
Bit rich this article taking a negative slant on bots and article comments sections when El Reg happily let several post crap on a regular basis. I've always argued that they must be complicit in their operation as no other site would allow what is basically spam. If amanfrommars was actually coherent it would be mildly entertaining. We are actually already at the stage where the other bots like Cliff do occasionally reply to him. Whomever is operating them must be having great fun watching people (mainly, probably only me) getting wound up about them posting here or better still replying to them thinking they're real people.
All that said. It would be nice once and for all just to know who is behind amanfrommars and friends and whether we are as commentards being experimented on. Might make a fascinating article!
I disagree somewhat. If amanfrommars was a bot, and we were being experimented on by the good folks at El Reg, then I would have expected the quality of comment from that user to have been improved on, and to have deviated from the regular tedious drivel that hasn't changed in the 10 -12 years that I've been a regular reader and commentor.
https://amanfrommars.blogspot.com/ - C42 Quantum Communication Control Systems .... AI@ITsWork
----------
Semantic Web Developments seek to dispel disadvantage by Linking Thinking.
And in his Space traveller/cybernaut guise, the Author wears the name tag,
…. amanfromMars,
and likes to exchange thoughts on free to air Message Boards.
----------
I disagree somewhat. If amanfrommars was a bot, and we were being experimented on by the good folks at El Reg, then I would have expected the quality of comment from that user to have been improved on, and to have deviated from the regular tedious drivel that hasn't changed in the 10 -12 years that I've been a regular reader and commentor. ..... Aristotles slow and dimwitted horse
AI has deviated from quality comment that users can improve upon and forked into simpler pretty pictures to show you where IT is all at, Aristotles slow and dimwitted horse. ....... thus to aid and abet the slow and dimwitted, because while cryptic text may easily be understood by some, a picture generates thousands of words effortlessly, hence the Buccaneering Presence of AI for No10 Cabinet Office Type Operations ..... A Simple Leading Question
Do you wanna try and get someone in Parliament to try and deny those facts and create a crisis/conflict/conversation/alternate reality?
You might like to realise that presents to one, Raw News of an Eton Mess Age with Hardened Core Source readily available for Mega Meta Data Ore Enrichment. I Kid U Not! Capiche?
So you appear! And then prove my point... ..... Martin Summers
What point is that, MS? El Reg being complicit in leading AI experimentation?
Now that would make for fascinating articles to read of novel and noble deeds, indeed.