By Debasish Mitra, VP Engineering, Mihup
For several decades, we have fancied self-driven vehicles that are operated through voice commands. Today, the evolution of AI and Natural Language Processing (NLP) technologies has helped us reach a point where conversational or voice AI has finally entered the realm of automobiles.Voice AI integrated vehicles of the day are a drastic shift from the initial usage of smartphone voice assistants such as Siri and Google as car navigation systems or for making phone calls. In the last ten years, there have been fascinating innovations in technology which have enabled the integration of voice AI with vehicles.
Many of the tasks in vehicles can be carried out through voice AI. Apart from making calls, the voice AI assistants in cars now offer options such as controlling temperature inside the vehicle, voice search and asking for recommendations related to restaurants, hotels, petrol pumps and other such facilities in the vicinity. Conventionally, a driver would use a text-based search on Google to find such information, something that could prove risky while driving. However, voice AI assistants eliminate the need for text inputs, and one can even lock/unlock the car/operate the sun roof, control infotainment systems, use navigation, connect with smart home devices and do online shopping using voice AI.
A major thrust towards adoption of voice AI as the technology that will be at the centre of mobility comes from smart vehicles.The world is rapidly moving towards electric vehicles and other smart as well as clean fuel mobility options. Tesla, Google and other leading tech companies are advancing towards introducing fully-autonomous vehicles on our roads. The more ambitious ones are even launching drone taxis. With such proliferation of AI-powered vehicles, voice AI is set to emerge as the touchless command centre for future transportation.
The reason why voice AI has entered the realm of reality from the sci-fi domain is that those working in the sector have managed some of the most impactful breakthroughs. Years ago, when the first voice assistants from American companies entered the scene, they were limited range tools trained in the US accent. Thus, they were mostly not able to understand the command when the speaker’s accent differed. Users would make multiple attempts to get tasks done, and even if the software eventually understood the command, it took much longer than anticipated. This inefficient and jarring experience made voice AI assistants more of a novelty and unsuitable for the in-vehicle ecosystem which demands a high level of accuracy as well as accent neutral response ability.
With the advancement in NLP, this problem is now being overcome by modern tech companies. In a country like India where numerous languages, dialects and accents are in vogue, the adoption of in-vehicle voice AI is being made possible by such innovators. Cloud-based voice AI tools have become a lot lighter, consume significantly low computes, and are also being conditioned to work in low-to-no-connectivity areas as well. This makes such systems perfect for integration with vehicles that might move around the countryside, in-and-out of network coverage areas. Further, the ability to respond to languages and dialects other than English ensures universal adoption and utility of such systems for all users. In Europe and Asia, there is a similar challenge of linguistic diversity which NLP based platforms can overcome.
In the future, shared mobility solutions are set to overtake private vehicle ownership, with people riding crewless vehicles that operate automatically. In such modes of transportation, voice will play the role of UI, and passengers/users will only speak to the vehicle instead of manually driving it or operating any feature. The first voice AI assisted vehicles are already out there on our roads.In another decade or so, it might be surprising to find a new vehicle that doesn’t talk back!