Photo: Soundhound | Edited
The automotive industry has employed voice assistance in their vehicles for many years, although the experienced hasn't impressed users. With the proliferation of Generative AI tools like ChatGPT, voice assistants are set to make a strong comeback in our vehicles.
The dream of talking to a computer has captivated people ever since this was a distant dream. Apple changed the landscape with Siri, and Google and Amazon followed with Voice Assistant and Alexa. Soon, the car industry, which secretly dreamed of beating the technology companies at their own game, was awash with half-baked voice assistants. Every car company wanted to develop its own voice assistant, claiming it would be better than Google Voice Assistant, Siri, or Alexa.
Their main problem was that people had a hard time talking to voice assistants in their cars. Even basic navigation features were hard to access via voice commands. "Sorry, I didn't get that" was a common reply, and people eventually gave up trying voice commands in their vehicles. The dream remained, though, as shown by a recent study by Soundhound, a voice AI company.
According to the study, half of the regular drivers said they are likely or very likely to use an in-car voice assistant powered by a Generative AI tool like ChatGPT. The likelihood for those that already use voice commands in their cars is even higher, with 75% willing to use an in-car voice assistant with Generative AI capabilities.
People want a voice assistant system that simply works for basic tasks such as navigation, radio, local search, making and answering calls, and text messaging. More sophisticated Generative AI functionality is wanted to answer questions related to vehicle maintenance, traffic rules, sights or attractions at the destination, recommendations for scenic routes, and other road trip/car-outing-related questions.
The study had a purpose: Soundhound is the maker of Chat AI for Automotive, which combines Generative AI capabilities with an established voice assistant. The result is a voice assistant that you can actually talk to, offering drivers and their passengers access to vast information resources with conversational capabilities. Soundhound also ensures their Chat AI avoids misleading responses common with large language models (LLM).
To avoid what is known as AI hallucinations, Soundhound applies a mix of software engineering and machine learning to select the correct responses from the most appropriate domain. They can be Generative AI-powered responses or answers that LLMs can't handle, such as real-time questions about navigation, weather, stocks, or flight status.
Depending on the car manufacturer's choice, Soundhound's Chat AI can integrate with any third-party Generative AI model to create a seamless conversational experience. Based on past events, we imagine carmakers scrambling to develop their own LLMs, which would be a disaster. Training an LLM is very compute-intensive and requires hardware and software resources that most automotive companies are unwilling or unable to mobilize.
Their main problem was that people had a hard time talking to voice assistants in their cars. Even basic navigation features were hard to access via voice commands. "Sorry, I didn't get that" was a common reply, and people eventually gave up trying voice commands in their vehicles. The dream remained, though, as shown by a recent study by Soundhound, a voice AI company.
According to the study, half of the regular drivers said they are likely or very likely to use an in-car voice assistant powered by a Generative AI tool like ChatGPT. The likelihood for those that already use voice commands in their cars is even higher, with 75% willing to use an in-car voice assistant with Generative AI capabilities.
People want a voice assistant system that simply works for basic tasks such as navigation, radio, local search, making and answering calls, and text messaging. More sophisticated Generative AI functionality is wanted to answer questions related to vehicle maintenance, traffic rules, sights or attractions at the destination, recommendations for scenic routes, and other road trip/car-outing-related questions.
The study had a purpose: Soundhound is the maker of Chat AI for Automotive, which combines Generative AI capabilities with an established voice assistant. The result is a voice assistant that you can actually talk to, offering drivers and their passengers access to vast information resources with conversational capabilities. Soundhound also ensures their Chat AI avoids misleading responses common with large language models (LLM).
To avoid what is known as AI hallucinations, Soundhound applies a mix of software engineering and machine learning to select the correct responses from the most appropriate domain. They can be Generative AI-powered responses or answers that LLMs can't handle, such as real-time questions about navigation, weather, stocks, or flight status.
Depending on the car manufacturer's choice, Soundhound's Chat AI can integrate with any third-party Generative AI model to create a seamless conversational experience. Based on past events, we imagine carmakers scrambling to develop their own LLMs, which would be a disaster. Training an LLM is very compute-intensive and requires hardware and software resources that most automotive companies are unwilling or unable to mobilize.
Google News