The perils of AI (Artificial Intelligence) in the PRC

April 17, 2023 @ 6:49 am · Filed by under Artificial intelligence, Computational linguistics, Language and politics

« previous post | next post »

Here at Language Log, for the last couple months, we've been having long, intense discussions about ChatGPT and other AI chatbots and LLM (Large Language Model) applications.  Now, it seems that the battle over such AI programs has reached the level of ideological warfare.

"America, China and a Crisis of Trust"

Opinion | The New York Times (4/14/23)

Indeed, a story making the rounds in Beijing is that many Chinese have begun using ChatGPT to do their ideology homework for the local Communist Party cell, so they don’t have to waste time on it.

I have some evidence that this might well be true.  Already about half-a-dozen years ago, my M.A. students from the PRC whose parents were CCP members told me that the government required daily interaction with the propaganda installed on their phones — upon pain of being demoted or dismissed.  They had to read a specified amount of Xi-speak and answer questions about the content.  This demanded a serious investment of time (hours).  It was considered to be especially onerous for those CCP members whose day jobs (doctors, bureaucrats, stock brokers, etc., etc.) already demanded a very full work schedule in the office.  So many, if not most of them, hired various human and electronic services to meet the obligations.

China state-run media China Daily claimed that ChatGPT "could provide a helping hand to the U.S. government in its spread of disinformation and its manipulation of global narratives for its own geopolitical interests." The Chinese government instructed Chinese tech companies not to offer access to ChatGPT services on their platforms.

(source)

The Chinese government blocks its citizens' access to vast amounts of information and technology available on the internet.  One can readily understand why such a totalitarian government would wish to forbid the free circulation of unfettered ChatGPT and other LLM-based bots:  they are garrulous, say what they wish, and are not programmed to self-censor, which is a sine qua non for "legal" participation on the internet behind the Great Firewall.

Remember the Bamboo Curtain?  It's still with us, just technologically transformed into the Great Firewall.

 

Selected readings

[Thanks to Bill Benzon]

April 17, 2023 @ 6:49 am · Filed by under Artificial intelligence, Computational linguistics, Language and politics


12 Comments

  1. Taylor, Philip said,

    April 17, 2023 @ 7:08 am

    I read recently that a British woman had avoided a fine/penalty after an alleged traffic offence by submitting to the court a letter of explanation/mitigation which had been drafted by ChatGPT at her request. So I suppose that despite my many misgivings, ChatGTP and its ilk might offer some benefits after all …

  2. KeithB said,

    April 17, 2023 @ 8:05 am

    Phillip Taylor:
    You would have to be very careful about this. For example, if you asked it for justification for not paying US income taxes (Coincidence: This is tax day in the US!) you could get garbage about admiralty flags, all-caps non-US persons and the fact that the Income Tax Amendment was not really ratified which is all over the internet.

  3. Peter Grubtal said,

    April 17, 2023 @ 9:30 am

    "ChatGPT to do their ideology homework for the local Communist Party cell"

    This sounds like the ideal occupation for ChatGPT.

  4. AntC said,

    April 17, 2023 @ 3:17 pm

    Or maybe not.

  5. Chester Draws said,

    April 17, 2023 @ 5:09 pm

    "… and are not programmed to self-censor, …"

    As shown several times already on LL, this simply isn't true. The range of topics isn't what the CCP wishes it to be, but there are plenty of topics where the current "AI" self-censors.

    (I call it "AI" because there is no intelligence there.)

  6. Victor Mair said,

    April 17, 2023 @ 9:51 pm

    =====

    Self-censorship is the act of censoring or classifying one's own discourse. This is done out of fear of, or deference to, the sensibilities or preferences (actual or perceived) of others and without overt pressure from any specific party or institution of authority. Self-censorship is often practiced by film producers, film directors, publishers, news anchors, journalists, musicians, and other kinds of authors including individuals who use social media.

    Article 19 of the Universal Declaration of Human Rights guarantees freedom of speech from all forms of censorship. Article 19 explicitly states that "everyone has the right to freedom of opinion and expression; this right includes freedom to hold opinions without interference and to seek, receive and impart information and ideas through any media and regardless of frontiers."

    (source)

    =====

    Self-censorship often takes place when one knows or believes something to be true or correct, yet intentionally and consciously withholds that information for fear of getting in political or other types of trouble. It is rampant in China, and is in addition to the heavy-handed, government imposed censorship from outside the individual (i.e., the self).

    Like Wikipedia, Google, and other disseminators of the free flow of information, ChatGPT and other LLM-based bots constitute a threat to the survival of totalitarian governments whose existence depends on the stifling of uninhibited information flow.

  7. Chester Draws said,

    April 18, 2023 @ 3:22 pm

    Like Wikipedia, Google, and other disseminators of the free flow of information, ChatGPT

    None of those have "a free flow of information". Sure, it isn't anything like the CCP censors, but they all three actively restrict certain information in accordance with their programmers' political beliefs.

    Wikipedia lock down pages to prevent free-editing, so you get the personal opinion of the editor. I've run across this regarding the Romanian-Hungarian war of 1919, a matter that still generates considerable heat.).

    Google rank certain sites higher than a free-running algorithm would. "Safe" climate change sites with almost no traffic are promoted over "unsafe" sites with considerable traffic.

    The AIs are much more restrictive. Any hot topic in modern life — Covid, Trump, race, gender, the range of human intelligence — is heavily censored. There's been too many examples of them racing off and giving the "wrong" answers to allow them free rein.

  8. liuyao said,

    April 18, 2023 @ 8:29 pm

    I don't dispute that these reports are factural, though the reality on the ground is that the public discourse is dominated by hype and excitement, and a sense of defeat (for the moment) for the Chinese tech giants.

    I was also worried that LLMs would be banned in China, as they "speak their mind" freely. But with the existing infrastructure of censorship in place, it's relatively easy to make the tech companies put the same screening on the output. These models are too big to run on your personal laptop or phone.

    The Chinese could turn it around and accuse American tech companies of imposing censorship, which they call "alignment". Tucker Carlson and Elon Musk are already making the case in the recent 60 Minutes interview.

  9. liuyao said,

    April 18, 2023 @ 9:02 pm

    Oops, I meant Carlson's interview, not 60 Minutes (which is about Google).

  10. Victor Mair said,

    April 18, 2023 @ 9:05 pm

    @Chester Draws:

    "…it isn't anything like the CCP censors…"

    QED

    It's a matter of degree, isn't it, like almost everything in the world, except perhaps temperature and a few physical things of that sort.

  11. KeithB said,

    April 19, 2023 @ 11:12 am

    Temperature isn't a matter of degree? Is that some sort of pun?

  12. Victor Mair said,

    April 19, 2023 @ 7:09 pm

    temperature can be absolute

RSS feed for comments on this post