Meta Description: AI-driven solutions are revolutionizing the way in which remote interpretation and translation services disseminate life-saving pandemic information across language barriers.
The COVID-19 crisis is a global pandemic and the novel coronavirus pays scant attention to national or linguistic boundaries. Even as nations and regions lockdown to stop the spread and “flatten the curve”, informing citizens what to do, those language gaps become an impediment to the dissemination of potentially life-saving information. Media, public officials, and healthcare providers struggle enough to get essential warnings and recommendations to populations which speak their own language. But what about the substantial minorities of residents who don’t know the dominant tongue? Are there solutions available for them? And what about the need to obtain important reports and studies from foreign language sources abroad? How can we expand and accelerate certified translation services so more people gain the knowledge they need to weather this storm of worldwide infection? Artificial Intelligence has a vital role to play
AI-Driven Translation Services and Tools to the Rescue
Those of us in the AI-savvy community are aware that, just five years ago, a machine language revolution took place. The introduction of neural network technology transformed the quality and efficiency of translation software. In 2014, neural nets were rarely and barely used for translation applications. By 2016, they were dominating the latest and greatest translation engines out there, scooping up the top awards at all major Artificial Intelligence competitions. The top tech companies jumped on the bandwagon, with Google, Facebook, Microsoft, Amazon, and Apple quickly jumping on the neural net bandwagon for their translation engines.
Five years later, these robot translators are giving human linguists a good run for their money. While expert translators still edge out software for complex translation tasks, machine translation beats average translator in blind tests scored by humans. Algorithms excel beyond flesh-and-blood linguists in churning out translations of structured content like weather reports, sports results, and financial updates. But all of this begs the question of the usefulness of these AI-driven tools to help humanity in communicating correctly about the scourge now imperiling the species. Statista estimates that in 2020, the worldwide market is $750 million, but the industry is expected to double in the next 4 years.
No Bot is Perfect: Where Machine Translation Falls Short
The AI-driven translation tools built into our smartphones, tablets, and personal computers take us a certain distance for internal research about the COVID-19 pandemic and the characteristics of the coronavirus. They can give the academic researcher or casual reader the gist of a foreign language news report or a paper from a scholarly journal. They can even handle the buzz-phrases that have recently entered the popular lexicon like “flattening the curve”: In English to French translation, Google Translate – widely considered the best translation app -- renders this correctly as “aplatir la courbe”. However, if one tries to render in French to English translation the acronym SAMU, short for France’s urgent medical service, Google’s translation engine is clueless.
In a 2019 study at the University of California at San Francisco, reported by Reuters, Google Translate translated correctly 594 out of 647 specialized medical sentences from Spanish to English. That may not sound too shabby at first, but it turns out that 29% of those errors were deemed capable of causing serious injury. Translating from Chinese resulted in even more mistakes (522 out of 647) and higher injury risk rates.
Companies and Governments Hustle to Translation Services to Meet the Need
In the current pandemic, then, it’s hardly surprising that companies and public health officials are not depending on machine translation for high profile translation assignments. Lives and reputations are on the line, so those who need to communicate in a foreign language prefer to rely on human experts for mission-critical assignments. It’s important to distinguish translation and interpretation, though the two terms are sometimes used interchangeably. Translation refers to written texts, while interpretation refers to spoken communications.
Language translation companies typically do both. Document translation services are their bread and butter, most of them delivering work online, usually turning around short assignments within 24 or 48 hours. More urgent are interpretation services, especially in the healthcare fields. Public officials and care providers in cities like New York, for example, are responsible for translating all important announcements and briefing, so that the substantial Hispanic population with LEP – Limited English Proficiency -- can understand what healthcare providers are telling them. The same applies to solutions for the hearing-impaired.
Both federal and state governments in the United States have mandated that medical information must by law be communicated in a language that patients understand. This requirement, legislated in the last two decades, has given rise to a robust industry for medical translation and interpretation.
AI Drives Over the Phone and Video Remote Interpretation for Urgent Care
AI-powered translation tools are stepping into the breach to overcome language barriers in healthcare. Any smartphone these days can deliver real-time voice translation as well as camera translation for rendering foreign languages text encountered in real-world settings.
Both Microsoft Translator and Google Translate offer these features as standard. Camera translation employs augmented reality to impose translated text over the original in a camera view or image. Voice Interpretation transcribes audio using a voice-to-text processor, renders and text-to-voice the machine translation. Because of their dependencies on machine translation, these features, while impressive, are still “rough around the edges” more useful for reading foreign language restaurant menus or street signs than complex medical reports or pharmaceutical prescriptions.
A more robust solution has emerged in recent years to combine AI-driven combined software-hardware solutions. Over the Phone Interpretation (OPI) has been around almost since the invention of the telephone. Lately, it has been overtaken by Video Remote Interpretation, delivered via videoconferencing tech (using solutions like Zoom and Skype) or dedicated video remote interpretation (VRI) systems offered by US-based companies like Stratus Video and Cyracom. Both providers provide combined hardware and software setups, then pair up with third-party translation agencies to deliver the interpreting in real-time to healthcare facilities, either via flat panel screens or smartphones and tablets controlled by physicians or patients. Their interpretation services are on-demand, with providers claiming connect times under a minute.
These remote solutions effectively replace the need for on-site translators and interpreters, a key consideration in the current circumstances where mobility is limited and the risk of infection from direct patient contact is high. The same holds true in the provision of translation and interpretation services in public health briefings and press conferences. The need for OPI and VRI, already combining for a $1.5 billion market in the US according to Statista, have been boosted by the urgencies of the pandemic. But for the foreseeable future, the language interpretation aspect of these solutions is likely to be delivered from a safe distance by humans, not AI-powered bots.
