Oscar-nominated film Her would have us believe that in the future we may fall in love with our operating systems. Or at the very least, it shows a much more fluid interaction with technology. Through mostly a speech interface, people are able to read email, search and view photos. Merge this with the Internet of Things that is taking over our present and you can see the power of APIs that speak your language.
If our future includes more wearables and a life surrounded by smart devices, voice is the killer interface. The first step to making this future work is turning spoken voice into something more easily understood. That means a voice-to-text API. When Siri was released four years ago, it used the Nuance Nina API to turn voice commands into text so its natural language processing engine could attempt to understand the meaning behind the words.
The APIs that really speak your language can convert it to someone else’s language. When Google shut down its translation API, the company had to know about the service’s popularity. The metrics-driven search giant likely saw the popularity as the biggest reason to close the doors on a service originally meant for researchers.
There are 79 translation APIs in the ProgrammableWeb directory.
Even Google knows that machine translation isn’t great for all uses. That’s why it uses the Gengo Human Translation API for offering YouTube captions in multiple languages. Gengo was the first of what is now at least eight human translation APIs.
Human translation may not fit the real-time nature of a world surrounded by smart things. Although, the network of multi-lingual individuals that must be built by human translation platforms could technically be leveraged in near-real-time. That network could help with the even larger problem of determining what the words were trying to communicate.
Indeed, semantic APIs may have the most promise in an age of vocal interfaces. Understanding intent is more important than simply translating text. Again, at Siri’s launch, the excitement was in its ability to infer meaning—”funny movies” means comedies.
Natural language processing APIs may be the answer to applications that can understand and react to meaning. You can find sentiment analysis and category recognition now, which go a long way toward interpreting our own language. Artificial intelligence isn’t quite at the point that’s displayed in movies, but when that time comes it’s reasonable to expect it will be powered not just by APIs, but by APIs that speak your language.