This is a link to a broad scope of problems I have been working on since 2005. I am asking about a particular case, not the set of zillions of topics many people work on.

I can't say I've achieved much but I have been interested in this question for some decades. I've given a talk: "Caring robots, more dangerous than killer robots" quite a few times in various places. 

When we ( Winfield et al 2010) produced the code of conduct for robotics researchers, I was the main one shouting about manipulating the vulnerable.

https://www.epsrc.ac.uk/research/ourportfolio/themes/engineering/activities/principlesofrobotics/

When we presented the final code a representative for a large computing company said "How can we make a profit if we can't to manipulate people?"  I've also had a disagreement with a major AI company over this issue (which legally I can't talk about). 

I'd be interested in a joint paper but should warn you that this one gets political very quickly. 

Thanks for both info and the advice!

RE: Do you know anyone working on ethical problems of AI being used for manipulation (both positive and negative)?

Robert Sparrow of Monash University has done work on ethical problems of AI and robots:  http://profiles.arts.monash.edu.au/robert-sparrow/?_ga=2.148694775.1742722190.1501030192-1799243749.1465776002

Thank you, Karl. I might be visiting Monash next month, precious!

Pedro, I am trying to find universals (also in beliefs) and I guess manipulation is one example, although libertarian paternalism might be considered as more positive in Japan than, let's say, in Poland. Is this what you want to say? Do you know anyone who works on this topic regarding technology?

https://www.theatlantic.com/magazine/archive/2017/09/when-silicon-valley-took-over-journalism/534195/

this is somewhat interesting and relevant to your question, in my opinion. "The pursuit of digital readership broke the New Republic—and an entire industry." 

Thanks for the article. It's not about automatic clickabait generation but shows one trend of how the manipulation could evolve when AI is used. 

"how the manipulation could evolve when AI is used..."  fair enough.  To that end, perhaps you might want to look into computer-aided Go game plan to the traditional 'ranking' system among professional players?  There is always an initial barrier to 'entrust' AI in fear of losing a sense of 'self'-hood?  It's more psychological than technological... in my opinion.

https://www.researchgate.net/post/Do_you_see_Googles_DeepMinds_AlphaGo_defeat_of_professional_9-dan_Go_player_as_good_as_IT_gets

Pushkin famously said: " translators are post-horses enlightenment" ... I had no clue what he possibly meant by 'post-horses', was it the way that he refrains from calling any individual 'Asses'  Arse from his POV? 

My answer, again, might not be what you were looking for... but this 1973 published translated essay had all the answers to your inquiry, in my humble opinion...

http://www.nybooks.com/articles/1973/08/09/translating-akhmatova/

@ S. 澍苍 Ye Laird

Post-horses were horses used to carry the post (i.e. mail). So post-horses delivered messages or information. Hence, translators were metaphorically post-horses in that they spread the ideas of the Enlightenment to other nations. Nothing wrong with that.

darn... I made the 指鹿为马 mistake again... haha...thanks for your correction, Prof. Pfeifer!  currently reading Andrei " Dissertation in the form of a Ballad" 

http://modernpoetryintranslation.com/poet/andrei-voznesensky/

https://www.theatlantic.com/magazine/archive/2017/09/how-america-lost-its-mind/534231/