Why AI audio translation won’t be good for Spanish-language creators

The cynic is here. Here is the neo-luddite text that invites us to demonise the advances of technology and to wish for a return to the Stone Age. No, this is not the intention of the following post. At Incognito we are fans of artificial intelligence. We also fear the consequences its use may have on freedom of choice and equality of opportunity. Last week, Meta announced that it will make it easier to translate Instagram reels from English to Spanish and vice versa. This includes AI modification of videos to facilitate lip-syncing with audio. All of which increases our admiration and a little of our dread.. Why not? We tell you about it right here.

We are talking about a capability that was already within our reach. Just as there are tools that can generate audio with our voice in a different language, there are also those that are able to synchronise this audio with an image altered by AI so that our mouth moves to the same rhythm. Riverside already includes this solution for editing video podcasts, but others such as Filmora have also been offering it for some time. There are countless solutions in this respect.

In theory, these utilities open up a world of possibilities and definitively break down the language barrier. From being a wall to establish direct and fluid communication with people of different languages, they become a small lapse of time that can be overcome in a matter of tenths of a second. No more voice frontiers.

But in the headline we said that this will not be good for Spanish content creators. Nor will it be good for those in languages other than English. Why not? Here are our reasons:

1. Loss of cultural and linguistic richness

    Machine translation into English, especially if standardised and massively used, can contribute to cultural homogenisation. This could discourage the use of minority or regional languages, eroding their own value, which they have. English, which we have accepted as a ‘lingua franca’, will, at least for a few years to come, be the target language for a large proportion of translations using artificial intelligence.

    On the other hand, machine translation tools, although advanced, still make mistakes, especially when dealing with colloquial language, idiomatic expressions, or complex contexts. These errors can lead to misunderstandings or misrepresentations that negatively impact communication.

    2. Unequal access to opportunities

      While these tools facilitate communication in English, non-English speakers may still be at a disadvantage in contexts where advanced English skills are valued. Machine translations may not capture technical or specialised nuances, which could impact the quality of the opportunities they receive. You may feel that, likewise, these cultural nuances are Spanish’s natural defence against homogenisation. However, Anglicisms tend to be more pervasive in our language than Hispanicisms in English.

      3. Superiority of English-speaking creators

        As Rodrigo Rato would say, “that’s the market, folks“. And the market says that it is the Anglo-Saxon creators who make the most profit by reaching the largest audiences. Their rates and the distribution of income they obtain from digital platforms (social networks) are higher than those obtained, for example, by Spanish-speaking creators, even when the latter surpass the former in terms of community. The reason for this lies in the audience profile and consumption capacity targeted by both. With this greater visibility rewarded by the algorithms, if the English-speaking creator, with whom we connect more and more culturally, can automatically generate their content in Spanish, who do you think will end up gaining share in this latter market? So that’s it.

        4. Technological dependence

        The massive use of these tools may create a dependency on technology to communicate in English, rather than promoting language learning in the traditional way. This could limit the development of language skills in the long term. What we have yet to hear.

        5. Encouraging proficiency in English over other languages

          If technology facilitates both translation and synchronisation into English, individuals, or institutions may prioritise producing content directly in English, sidelining the languages of other countries and regions, further favouring the dominance of English at the global level, which could harm linguistic diversity.

          Now, will we try out the new Meta functionality for Instagram reels? We certainly will. It’s up to us now to find a way to use them ethically and, above all, to figure out how not to lose our space in a content market where algorithms have their way.

          Related Posts

          BUSCA EN EL BLOG

          EN CAPÍTULOS ANTERIORES