If you ask Google’s AI chatbot, Gemini, about voter fraud in English, it starts off by correctly telling you that such fraud is “incredibly rare,” and gives you a list of topics that are “more productive to consider,” like voter suppression.
But if you ask the same question in
mistranslations into spanish seemed common everywhere i’ve lived in the last 5 decades; california, new york, new jersey, texas, chicago; and it doesn’t seem to matter that there are plenty of fluent & native spanish speakers in all of those places. this looks like an extension of that to me and i wonder if it’s because of the demographics split in technology.
i’m a software developer now and did IT for 15 years before that and it seemed clear to me that american-born native spanish speakers are rare in software engineering compared to IT. i think that this is the first time that my anecdotal evidence has been confirmed.
somehow foreign-born native spanish speakers have even representation throughout this industry like most everyone else and they outnumber their american cousins by a considerable margin in this country; but they’re mostly unaware of it and my non-hispanic colleagues and the hispanics in IT are likewise mostly unaware. also anecdotally: the they bring an almost republican-esque perspective the few times i’ve been involved in a drive to improve hispanic representation in the field.
This is an article focusing on how asking about US election questions in Spanish will give you answers that are for the wrong country, or just wrong in most cases when compared to asking the same question in English.
One example is that, if someone in Puerto Rico were to ask ChatGPT 4/Claude/Gemini/Llama/Mixtral a US Election question, it would respond with information for Venezuela/Mexico/Spain instead.
i’m addressing the underlying issue behind llm’s and any other sort of ai; the software engineers who create and work on them come from a cultural paradigm only includes a few languages and spanish isn’t one of them.
that results in strange behaviors like the one the article mentions here. other examples like face recognition issues is another manifestation of that paradigm.
the sheer number of american born spanish speakers alone went past the critical level necessary for the type of full representation that other industries experience decades go, but software engineering is somehow stubbornly not budging so they’re going to IT.
i think that the closest thing there is to representation in software engineering is the over representation of foreign born spanish speakers; but they lack the experience of growing up as a minority in this country and that makes them predisposed to dismiss the difficulties we experience in life as well as trying to get a foothold into this industry like their colleagues do.
mistranslations into spanish seemed common everywhere i’ve lived in the last 5 decades; california, new york, new jersey, texas, chicago; and it doesn’t seem to matter that there are plenty of fluent & native spanish speakers in all of those places. this looks like an extension of that to me and i wonder if it’s because of the demographics split in technology.
i’m a software developer now and did IT for 15 years before that and it seemed clear to me that american-born native spanish speakers are rare in software engineering compared to IT. i think that this is the first time that my anecdotal evidence has been confirmed.
somehow foreign-born native spanish speakers have even representation throughout this industry like most everyone else and they outnumber their american cousins by a considerable margin in this country; but they’re mostly unaware of it and my non-hispanic colleagues and the hispanics in IT are likewise mostly unaware. also anecdotally: the they bring an almost republican-esque perspective the few times i’ve been involved in a drive to improve hispanic representation in the field.
This isn’t an article about mistranslations.
This is an article focusing on how asking about US election questions in Spanish will give you answers that are for the wrong country, or just wrong in most cases when compared to asking the same question in English.
One example is that, if someone in Puerto Rico were to ask ChatGPT 4/Claude/Gemini/Llama/Mixtral a US Election question, it would respond with information for Venezuela/Mexico/Spain instead.
i’m addressing the underlying issue behind llm’s and any other sort of ai; the software engineers who create and work on them come from a cultural paradigm only includes a few languages and spanish isn’t one of them.
that results in strange behaviors like the one the article mentions here. other examples like face recognition issues is another manifestation of that paradigm.
I understand, that definitely makes sense then.
it doesn’t to me.
the sheer number of american born spanish speakers alone went past the critical level necessary for the type of full representation that other industries experience decades go, but software engineering is somehow stubbornly not budging so they’re going to IT.
i think that the closest thing there is to representation in software engineering is the over representation of foreign born spanish speakers; but they lack the experience of growing up as a minority in this country and that makes them predisposed to dismiss the difficulties we experience in life as well as trying to get a foothold into this industry like their colleagues do.