Some red flags around Google’s language-to-text AR glasses

By S.A. Applin 3 minute Read

​​Google seems to be engaged on AR glasses as soon as once more, however this time, it’s exhibiting a brand new function that interprets speech to readable textual content. 

At final week’s Google I/O 2022, the corporate demonstrated an AR glasses prototype that may translate spoken language into readable show textual content. Google has not hinted whether or not they’re creating these as a product, or when, however the truth that they confirmed them to builders is indicative that they’re pondering of the best way to prolong the mannequin of AR glasses to make the most of their gigantic datasets and present applied sciences.

If Google strikes ahead with the product, it’s seemingly that it’s going to body it as a tool that might  try to interrupt down language limitations. Sounds nice, proper? No extra looking for Google Translate on the internet and pecking phrases into our cell phones to translate issues. When (or if) these hit the market, we’ll lastly have the ability to learn international indicators, order appropriately in eating places, and even make new buddies extra simply once we journey. More considerably, there can be a option to shortly translate communication within the occasion of an emergency, when folks might not all communicate the identical language. On one other stage, these “translation glasses” may additionally open up communication channels for the deaf and laborious of listening to group, giving them a brand new option to talk with these round them.

However, as with all new expertise concepts, Google’s translation glasses may include an enormous social price: to our privateness, our well-being, and our cooperation with one another in our communities. What does it imply when Google turns into the translator for our lives, and are we snug with that notion? 

The downside with any sort of expertise translation gadget is that it has to “listen” to these round it, to obtain the info to translate. And if the AR glasses are listening, we might want to know what, or whom, they’re listening to—and when they’re listening. At the second, we don’t know if these glasses will have the ability to distinguish between multiple individual at a time, both. Also, we might want to know whether it is authorized for these glasses to hear with out consent—and if one wants the consent of somebody to file them in an effort to translate them, will one want the glasses to translate the consent? We don’t know if sooner or later, these glasses can have the capability to file what they translate, nor will we all know if they may establish whom they’re recording at any given time, or inside what vary they’re able to listening. And If they’re recording glasses, and even with the transcribed textual content, we’ll must know if  that’s saved someplace that may be erased, and if folks may opt-out in a public house with out being recorded whereas doing so.

Let’s assume for the second that these Google glasses gained’t file us, and that Google manages to determine consent and permission. Given that, in our crowded, noisy world, the same old issues with speech to textual content may nonetheless abound within the type of misunderstandings, misrepresentations, and so forth., in what Google ‘hears’ and what it writes on account of that listening to. The tech may additionally have quite a lot of misspellings and confusion with mixing languages. As The Verge identified, many people “code switch” utilizing phrases from many various languages interspersed, with the added complexity of not all of them studying from left-to-right, which can must be accommodated, too. 

Now add to that an combination inhabitants utilizing these whereas wandering round, which invokes a lot of what I wrote with Dr. Catherine Flick about Meta’s pre-Ray-Ban Stories Project Aria glasses. Many of the identical points persist, aside from that with these new Google glasses, folks could also be strolling round and studying transcripts, which once more, is extra like what was happening within the early days of cell telephones and Divided Attention, creating doubtlessly harmful outcomes as distracted folks stroll into site visitors or fall into fountains.

One of the principle issues with the glasses is Google’s seeming assumption right here that expertise can resolve cultural issues—and that if the expertise isn’t working, the answer is to develop and apply extra expertise. In this case, fixing cross-cultural communication issues can’t be absolutely solved with language translation. Tech may help, however these glasses gained’t translate tradition or cultural norms resembling whether or not somebody is snug being direct or oblique, or any one among multitudes of cultural nuances and cues discovered within the ways in which completely different folks in several teams talk with one another. For that, we’d like different people to information us.

S.A. Applin, PhD, is an anthropologist whose analysis explores the domains of human company, algorithms, AI, and automation within the context of social programs and sociability. You can discover extra at @anthropunk and

Source hyperlink

Leave a Reply

Your email address will not be published.