ChatGPT is popular: in-vehicle voice interaction is heading towards the mainstream

Publisher:科技狂人Latest update time:2023-02-22 Source: AutoR智驾 Reading articles on mobile phones Scan QR code
Read articles on your mobile phone anytime, anywhere

Share For the automotive field, as current new energy vehicles have increasingly intelligent attributes, voice control in particular has become a standard feature of some high-end smart models. With the support of ChatGPT technology, will the in-vehicle voice track usher in an opportunity for rapid development?


In November last year, ChatGPT was born, and it became a phenomenon-level application in just a few months.


Compared with the metaverse where "everything can be put into it", this chatbot seems to be more down-to-earth and more practical. After ChatGPT was launched, many users talked to it out of curiosity, and its answers to many big questions were truly amazing.


But beyond the surprise, people soon discovered that most AIGC companies are currently in the C-side diversion stage, and everyone is still exploring what kind of business model they will use to make money in the future. At this time, only by finding suitable application scenarios for the chatbot can its commercial value be released.


For the automotive field, as current new energy vehicles have increasingly intelligent attributes, voice control in particular has become a standard feature of some high-end smart models. With the support of ChatGPT technology, will the in-vehicle voice track usher in an opportunity for rapid development?


A burst of warm air in the car voice industry


In-vehicle voice interaction system is a subdivision of smart cars. Compared with traditional touch screen operation and button operation, vehicle control is completed through voice, and the driver does not need to leave the steering wheel with his hands; compared with gesture operation, the learning cost of voice interaction is also lower.


It is precisely because of these advantages that in-vehicle voice interaction has been popular in the market in the past few years, and it was once an intelligent configuration that car companies rushed to launch. According to Tianyancha APP search, the number of domestic companies related to vehicle voice has reached 196.


However, since last year, the pre-installed in-car voice market has entered a bottleneck period. In particular, the room for increase in the pre-installed installation rate has been limited, and product prices have continued to decline. This market segment has also become much more deserted than before. Some suppliers of in-car products The sector's revenue even declined.


Take iFlytek as an example. As the leader in China's intelligent voice industry, according to its 2022 semi-annual report, automotive intelligent network-related businesses (mainly from in-vehicle voice) achieved revenue of 169 million yuan, a year-on-year decrease of 5.72%.


As we all know, intelligence is a major trend in automobile development. Why has the in-vehicle voice industry failed to grow along with the trend?


Judging from user feedback, many in-vehicle voice systems are generally weak in intelligence and have an unsatisfactory experience. In this case, only a few high-end brands such as Xiaopeng and NIO will use more advanced in-car voice systems as a differentiated selling point.


"The installation rate of in-vehicle voice is actually not low, and the price is generally not expensive, but the user experience is very average and the gimmick component is relatively large. In the eyes of many users, it is just a dispensable 'tasteless' configuration." A certain in-vehicle voice manufacturer The sales manager told Tan Qing about AI representation.


In fact, some car companies' mastery of voice interaction technology is not mature enough, resulting in poor product experience. Not only cannot they improve the smart experience, but they also cause trouble to users.


For example, a car influencer previously exposed a weird bug in Ji Krypton 001: the voice control in the rear row falsely alarms once every 5 minutes on average, mistakenly believing that someone in the back row is calling it. "If you want to listen to music and rest quietly in the car, you have to hold your breath and stay still. Don't let the car detect that you are still alive, otherwise it will [I am in the back row]."


It can be found that the current car voice industry is mainly stuck in immature technology and tasteless experience. So with the support of ChatGPT's large model, self-learning and other technologies, if the application experience of car voice can achieve a breakthrough, it may really be able to revive the tepid car voice industry.


In fact, some car companies and suppliers have already started to take action.


Recently, Jidu Auto, which is backed by Baidu, has announced that it will integrate the comprehensive capabilities of Wen Xinyiyan, the "Chinese version of ChatGPT", to create a large model for smart car scenarios to enhance the natural performance of its first model ROBO-01 car robot. comminicate.


iFlytek, as the leading supplier in the automotive voice industry, also sees opportunities in this wave. Recently, iFlytek Research Institute of HKUST released a 10,000-word interpretation of ChatGPT and revealed its technology accumulation and related industry layout. It is reported that iFlytek plans to launch a "1+N" architecture to improve the practicality of large models in fields such as automotive human-computer interaction.


The current arms race around conversational robots has begun. For car companies, perhaps they should think about how to deeply integrate the new generation of in-vehicle voice with in-car scenes, cultivate the habit of using voice interaction among drivers or passengers, and even form an effective business model.


As chatbots “get on board”, emotional intelligence is no longer a gimmick?


Many car companies and suppliers are in the process of exploring how conversational robots can empower in-vehicle voice. For car companies, once the technology combination of conversational robots and in-vehicle voice interaction matures, car brands that focus on emotional intelligence may be able to discover more dazzling features in the car experience.


The cruelty of the car-making track is that there are not many accurate and effective brand positionings. For example, Lantu Motors, a subsidiary of Dongfeng, is positioned as a "zero-anxiety high-end smart electric brand", but zero-anxiety is only because it has extended-range models, and its smart configuration is also weaker than that of similar competing products such as Xpeng and Weilai. The luxurious interior can no longer be seen as high-end.


In addition, there are Skyworth Motors, which focuses on "health", and Avita, which focuses on "emotional intelligence". Judging from their weak sales, their brand positioning has failed to impress users and only serves to highlight that they are different from other brands. Judging from the overall evaluation of many media on the market, the brand positioning of the above car companies is actually "gimmicks greater than reality."


For many new forces that have not yet mass-produced the "life and death line", positioning should neither be too low nor live up to the name. This is a difficult problem.


In the era of new energy revolution, new powerful car companies have emerged, resulting in too many car brands and smart cars falling into homogeneous competition. How to find a good brand positioning is crucial for a car brand to break out of the circle.


But the difficulty is that good positioning has been seized by leading car companies, and these car companies will continue to invest in R&D and brand operations around brand positioning, thereby strengthening the label of brand positioning and building competition barriers for enterprises.


With the catalysis of the ChatGPT craze and the layout of relevant suppliers, start-up car companies focusing on emotional intelligence may find a new solution to brand positioning.


From the perspective of technical classification, conversational robots are divided into task type (Task) and chat type (Chat). The current in-vehicle voice function is useless because its bottom layer is an incomplete task-based dialogue system. Once a chat-type dialogue system is introduced, the car-machine system may open the door to a new world and shift towards more personalized and emotional voice interaction.


At that time, smart cars, as a new generation of mobile terminals, will also become personal smart life assistants and even virtual friends.


Judging from the IQ displayed by ChatGPT, when this kind of dialogue robot based on large models is constantly "fed" by big data, it is already capable of chatting with people, and is knowledgeable enough, and even has high emotional intelligence. In other words, conversational robots can make the intelligence of in-vehicle voice assistants closer to real people. With the support of this technology, emotional intelligence may become a good brand positioning for car companies.


When in-vehicle voice has achieved a leap from "performing tasks" to "accompanying and chatting", it may become possible to monetize traffic.


Weilai’s advertisements on the car and machine have previously caused dissatisfaction among some users. In fact, this matter can be understood in the abstract as the car and machine are not intelligent enough. Users can only accept it to receive and follow the instructions of pedestrians, but not accept it in turn. People output information.


In other words, the intelligence of such a car is too weak to escape the essence of tool attributes. Trust in cars and machines is too low. If you push ads at this time, users will instinctively resist.


And once the car-mounted AI voice assistant evolves into a chat-type robot, after interacting with the user for a long time, it will understand the user like a tapeworm in the user's stomach. It may even become inseparable from the user, and you will habitually ask it if you have any questions. , who would resist if you recommend something to the user at this time?


Last year, New Oriental transformed itself into a live broadcast to sell goods. The anchor Dong Yuhui relied on his rich knowledge reserve and blurted out poems and verses, leading the trend of "knowledge-based sales". Many viewers go to the Oriental Selection live broadcast room to buy the same thing, which actually highlights the commercial value of an "intellectual chat friend".

[1] [2]
Reference address:ChatGPT is popular: in-vehicle voice interaction is heading towards the mainstream

Previous article:A new generation of Apple CarPlay will be released at the end of the year!
Next article:Research on in-vehicle information services 1: The control range is expected to be expanded to the entire vehicle, and cockpit games will become the next direction

Latest Automotive Electronics Articles
Change More Related Popular Components

EEWorld
subscription
account

EEWorld
service
account

Automotive
development
circle

About Us Customer Service Contact Information Datasheet Sitemap LatestNews


Room 1530, 15th Floor, Building B, No.18 Zhongguancun Street, Haidian District, Beijing, Postal Code: 100190 China Telephone: 008610 8235 0740

Copyright © 2005-2024 EEWORLD.com.cn, Inc. All rights reserved 京ICP证060456号 京ICP备10001474号-1 电信业务审批[2006]字第258号函 京公网安备 11010802033920号