Microsoft Strives to Make Star Trek’s Universal Translator a Reality

Microsoft Strives to Make Star Trek's Universal Translator a Reality



Key Takeaways

  • Microsoft is developing AI-powered real-time translation for Teams by 2025, revolutionizing virtual communication.
  • Challenges with real-time translation technology paralleled those depicted in
    Star Trek: Enterprise
    .
  • If Microsoft’s Interpreter works as described, it will make Star Trek’s Universal Translator concept a reality.

The Star Trek franchise has inspired dozens of technological advances in the past five decades. To some extent, cell phones, tablets, voice-activated computers, video conferencing and more all exist because of Star Trek. Now, technology giant Microsoft is tackling another piece of Trek technology — the Universal Translator.

The Universal Translator is one of the most important pieces of technology in the Star Trek universe. This technology, which is built into Starfleet’s communicators, provides real-time translation of alien languages. Since verbal communication is the primary way beings interact with each other, the premise of Star Trek — making contact and developing cultural relationships with new alien species — would be virtually impossible without the Universal Translator.

Related
Star Trek’s Translator Technology, Explained

Exploring how the google translate of the future works.

In today’s world, where people across the world are connected through social and business relationships, real-time translation technology is more important than ever. Though there have been significant challenges in developing this kind of technology, Microsoft just announced that AI-powered real-time translation services will be available in Microsoft Teams in early 2025.

Microsoft’s Interpreter

According to Microsoft’s November 19th press release, upgrades to its existing Copilot Studio coming in 2025 will include incredible enhancements to the current translation capabilities in the Microsoft Suite of products. Microsoft claims that Interpreter will be able to translate anything in a virtual meeting held on Microsoft Teams in real-time. No more waiting for an interpreter as a go-between. No more pausing meetings to wait for the translation software to catch up. As someone speaks in a meeting, the person on the other end will hear their speech in their native language.

The program draws on a massive AI data bank to both interpret the speaker’s words and translate them into the listener’s native language. Microsoft even says that Interpreter will have a setting to change the sound of the speaker’s voice so it sounds more like the listener’s native speech, making it easier to understand.

The demo of Interpreter is at the 45-second mark in the video above.

The Challenges of Real-Time Translation

Hoshi Universal Translator
Screen grab from Star Trek: Enterprise episode “Broken Bow”

Long before AI was advanced enough to provide real-time translation services, Star Trek: Enterprise predicted some of the challenges tech companies would face in building a Universal Translator.

As explained in a study about the accuracy of AI translation programs published in iScience, translation programs can only work when they have accurate information about both the foreign and native languages. Humans have to train the translation programs, fix their errors, and update the programming with new information.

In Enterprise, early versions of the Universal Translator heavily relied on a human translator. When the Universal Translator couldn’t decode an alien language, Ensign Hoshi Sato, a gifted linguist, was responsible for filling in the gaps and reprogramming the Universal Translator. However, Sato was working with completely alien languages, and often, she struggled to learn the grammar, syntax, and nuances of these languages, interpret them, and reprogram the Universal Translator.

Though they’re not working with alien languages, people working on real-life translation programs have run into similar issues, especially when trying to create real-time translation programs. Each language has its own quirks, idioms and colloquialisms, as well as multiple accents and dialects. To translate all of those properly, and convey the ideas within the right context, translation programs needed to understand all those details, recognize them, and interpret them in real-time. To achieve this, translation programs need to analyze an immense amount of data, and learn from it.

Until recently, AI wasn’t advanced enough to aggregate all this data and organize it in a way that made it accessible to translation programs. But the amazing advances in AI technology that have happened over the past few years have allowed companies like Microsoft to rely on AI translation as the basis for programs like Interpreter.

If Microsoft’s Copilot and Interpreter actually work the way the company claims, it will truly revolutionize the way people virtually communicate, and Star Trek’s dream of a Universal Translator will become a reality.

Star Trek_ Enterprise

Star Trek: Enterprise

Release Date
September 26, 2001

Seasons
4

Creator
Rick Berman, Brannon Braga

Number of Episodes
98

Network
UPN

Sources: Microsoft, iScienceStar Trek: Enterprise

Source link