Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AI, Wikipedia, and uncorrected machine translations of vulnerable languages

technologyreview.com

Post reply on HN
Print view View original

61–70 of 84 posts

Re: AI, Wikipedia, and uncorrected machine translations of vulnerable languages

#61

Post by baobabKoodaa »

tdeck wrote: post #46
aucisson_masque wrote: post #6

> Wehr, who now teaches Greenlandic in Denmark, speculates that perhaps only one or two Greenlanders had ever contributed. That's the core issue, it's not those who use AI translator or worst like Google translate. If there isn't any Greenlander to contribute to their Wikipedia, they don't deserve to have one and instead must rely on other languages. The difference between an empty Wikipedia and one filled with trans…

It is worse. Imagine if you were trying to learn English from this phrasebook, written by someone who didn't speak English: https://www.exclassics.com/espoke/espkpdf.pdf Wikipedia is prominent. Wikipedia articles in a language without much representation become prime examples of that language to those who read them.

Jesus. What did I just read?

Re: AI, Wikipedia, and uncorrected machine translations of vulnerable languages

#62

Post by tdeck »

baobabKoodaa wrote: post #61
tdeck wrote: post #46

Earlier quoted context omitted.

It is worse. Imagine if you were trying to learn English from this phrasebook, written by someone who didn't speak English: https://www.exclassics.com/espoke/espkpdf.pdf Wikipedia is prominent. Wikipedia articles in a language without much representation become prime examples of that language to those who read them.

Jesus. What did I just read?

https://en.wikipedia.org/wiki/English_as_She_Is_Spoke

Re: AI, Wikipedia, and uncorrected machine translations of vulnerable languages

#63

Post by mig39 »

debtta wrote: post #47
simonw wrote: post #4

I'm surprised this story didn't mention the scandal with Scots Wikipedia: https://www.theguardian.com/uk-news/2020/aug/26/shock-an-aw-... > an American teenager – who does not speak Scots, the language of Robert Burns – has been revealed as responsible for almost half of the entries on the Scots language version of Wikipedia It wasn't malicious either, it was someone who started editing Wikipedia at 12 and naively fa…

The background here is that Scots is not really a language. Try asking a Glasgow taxi driver who addresses you in 'Scots' whether he knows any English. Robert Burns wrote in English, with some of his spelling reflecting pronunciation in the Scottish English dialect. The people who want it to be considered as a language for political reasons cannot be bothered to translate Wikipedia themselves. They read and edit Engl…

> The background here is that Scots is not really a language.

This is supremely ignorant. Scots is its own language. It's a 'brother' or 'sister' of English, with both English and Scots being descendants of West Germanic languages.

The fact that many (all?) Scots speakers also speak English doesn't mean Scots not a language on its own.

You could make your exact same arguments that Irish isn't a language because you could ask a Cork taxi driver whether he knows any English.

Scots = a language with some of the same ancestors as English.

Scottish English = a dialect (and accent) of English

Scots Gaelic = another language, with the same ancestors as Irish and Manx.

Re: AI, Wikipedia, and uncorrected machine translations of vulnerable languages

#64

Post by muldvarp »

strogonoff wrote: post #18

Wikipedia editors is among the many communities that have for a long time mostly successfully relied on the tendency of relatively superficial, easy to validate capabilities (such as being able to use a website, write something resembling real language, and handle basic communication) to correlate with more valuable but harder to validate qualities (such as ability to write reasonably well and follow rules/guidelines…

LLMs destroying Wikipedia would be incredibly sad and is one of the things that makes me think that LLMs will have a strong negative impact on the lives of most people.

Re: AI, Wikipedia, and uncorrected machine translations of vulnerable languages

#65

Post by Xss3 »

mikemcquaid wrote: post #59
debtta wrote: post #47

Earlier quoted context omitted.

The background here is that Scots is not really a language. Try asking a Glasgow taxi driver who addresses you in 'Scots' whether he knows any English. Robert Burns wrote in English, with some of his spelling reflecting pronunciation in the Scottish English dialect. The people who want it to be considered as a language for political reasons cannot be bothered to translate Wikipedia themselves. They read and edit Engl…

Sort of? The Glaswegian taxi driver may not consider themself to be speaking a different language but, if speaking to another local and leaving aside pronunciation, they’d use words, phrases and even grammar that’s incomprehensible to someone with no experience with Scots. I’m a “posh Scot”, raised middle class in Edinburgh so my accent is minimal and thickens up or softens depending on who I’m speaking to. Even for…

Brits do the same. At best it is a dialect at worst an accent. A lot of (most of) Scots is still English but spoken with different grammar or unfamiliar phrases and unfamiliar pronunciation.

Sort of like extreme cockney rhyming slang or for a more modern example thick BME* full of slang.

* = British Multicultural English, think fam n blud, lots of Jamaican english influence plus south east asian influence.

Re: AI, Wikipedia, and uncorrected machine translations of vulnerable languages

#66

Post by georgefrowny »

muldvarp wrote: post #64
strogonoff wrote: post #18

Wikipedia editors is among the many communities that have for a long time mostly successfully relied on the tendency of relatively superficial, easy to validate capabilities (such as being able to use a website, write something resembling real language, and handle basic communication) to correlate with more valuable but harder to validate qualities (such as ability to write reasonably well and follow rules/guidelines…

LLMs destroying Wikipedia would be incredibly sad and is one of the things that makes me think that LLMs will have a strong negative impact on the lives of most people.

I have previously translated a very small handful of redlink articles into English from another language. Chasing down the sources in the other language and synthesising and cross-referencing with English sources is a fun challenge. To the best of my knowledge, I did an OK job.

While translation tools are a godsend for that, as well as life in general when dealing with a language I am not that good at, LLMs make me increasingly reluctant to do that much more because there is no way I could detect AI slop in a second language. For all I know I'd be translating junk into English and enabling translingual citogenesis.

Bad as the slopwave is for native speakers, it's absolutely brutal for non-native speakers when you can't pick up on the tells. Maybe the gap will narrow and narrow until the slop is stylistically imperceptible.

Re: AI, Wikipedia, and uncorrected machine translations of vulnerable languages

#67

Post by overfeed »

debtta wrote: post #47
simonw wrote: post #4

I'm surprised this story didn't mention the scandal with Scots Wikipedia: https://www.theguardian.com/uk-news/2020/aug/26/shock-an-aw-... > an American teenager – who does not speak Scots, the language of Robert Burns – has been revealed as responsible for almost half of the entries on the Scots language version of Wikipedia It wasn't malicious either, it was someone who started editing Wikipedia at 12 and naively fa…

The background here is that Scots is not really a language. Try asking a Glasgow taxi driver who addresses you in 'Scots' whether he knows any English. Robert Burns wrote in English, with some of his spelling reflecting pronunciation in the Scottish English dialect. The people who want it to be considered as a language for political reasons cannot be bothered to translate Wikipedia themselves. They read and edit Engl…

What counts as a language is almost always determined by "political reasons" - as the witticism goes: "A language is a dialect with an army and navy."

There exists dialects that are less mutually intelligible than apparently distinct languages, and the designation of each as "dialect" or "language" is political. Language is often a proxy for culture, and political actors may wish to suppress or boost the legitimacy of such cultural expression depending on their aims.

Re: AI, Wikipedia, and uncorrected machine translations of vulnerable languages

#68

Post by vacuity »

Uehreka wrote: post #8
ratg13 wrote: post #3

It's ironic that the "solution" to the problem is being driven by yet another person that isn't native to Greenland. While they may be a Greenlandic teacher, it's almost assured that they are teaching western Greenlandic, which is similar to Canadian Inuktitut. People in the East of Greenland speak a language that has similarities, but is different enough in vocabulary and sounds that it's often considered a separate…

I wish people on HN would stop acting like “magic arbiter” solutions are “not realistic”, when in reality it’s the only way things have every worked. Are federal judges “magic arbiters”? Yes. Do judges make bad calls? Yes. Do we not like when large numbers of judges who are unfriendly to our side get life appointments? Yes. Has anyone proposed an actual better way of solving these kinds of problems? No. So to get bac…

> Has anyone proposed an actual better way of solving these kinds of problems? No.

This is the heart of the matter. Nothing is good or bad in a vacuum, but when two things (say, outcomes) can be compared, distintions can be drawn. Noticing flaws in the present can't be contrasted with simple models of "the better solution"; this is comparing apples to oranges. Address both the good and the bad of the present, including the days where nothing noteworthy happens and therefore below the awareness of most people, and the good and the bad of an elaborated counterpart.

Re: AI, Wikipedia, and uncorrected machine translations of vulnerable languages

#69

Post by bawolff »

thaumasiotes wrote: post #40
bawolff wrote: post #17

Earlier quoted context omitted.

> People in the East of Greenland speak a language that has similarities, but is different enough in vocabulary and sounds that it's often considered a separate language and not a dialect. If this is true, then the easy solution would be to just have two separate wikipedia editions (assuming there is interest). After all if we have en, sco, jam and ang, surely there is room for two greenlandics. The limitting factor…

> the easy solution would be to just have two separate wikipedia editions (assuming there is interest) That's... a reach. An easier , and much more realistic, solution would be to just have one edition in Danish, which was already noted as the language Greenlanders have in common.

Well, is the point to have greenlanders be able to read it, or is it to preserve a dying language?

Re: AI, Wikipedia, and uncorrected machine translations of vulnerable languages

#70

Post by internet_points »

bradley13 wrote: post #44

I've lived in a couple of countries where there is a "vulnerable" language. I understand the emotional attachment that the native speakers have to their language. However, in the larger picture: languages evolve. New ones develop, old ones die. Do artificial attempts to "rescue" a language really make sense?

The AI slopwave is about as close to natural linguistic evolution as world war 2 was to natural selection (..aaand there we hit godwin's law, I'll see myself out)
Post reply on HN