Since the beginning of the year, ChatGPT has become popular, and smooth chat, drawing, script creation, and program code writing are all not a problem, and it can be called a "mind-reading" intelligent assistant.
*Image source: Weibo
More representative is this year's Shanghai Auto Show, after the BMW ice cream incident, netizens thought that the official apology letter was bad, and then let ChatGPT rewrite a copy, and the result was rave reviews.
From Yu Chengdong to Li Xiang, they all said that they would let the AI big model get on the car.
This is not Mercedes fighting, it is estimated that it is also urgent!
In the past two days, it was announced that the vehicle-computer interaction system MBUX was connected to ChatGPT, which is also its first access to ChatGPT.
After access, the Mercedes-Benz voice assistant will not only have a smoother conversation, but also help users solve problems such as what to eat for dinner, and even tell jokes and other emotional resonance.
It is expected that around 900,000 car owners will be able to enjoy the feature.
The old fox was conquered by the speed of Mercedes this time!
Admittedly, cars are getting smarter as more and more big-model AI gets on board, but do we really need it?
What problems does artificial intelligence solve in the car?
First, what problems do you solve in the user experience?
Old Fox believes that the fundamental core lies in the change of use scenarios.
Compared to traditional cars, which simply go from point A to point B, our main needs are safety, economy and comfort.
But under the definition of artificial intelligence, the car will eventually move towards L5 autonomous driving, basically do not need a person to take over the vehicle, we have more free time in the car, the car is also defined as a third space outside the home and company.
Therefore, leaving aside complicated technical theories, intelligent cockpits and intelligent driving systems have become the core of consumers' easiest-to-perceive.
The former is represented by the ideal "color TV and refrigerator";
The latter is a breakthrough in Huawei and Tesla FSD that does not rely on high-definition maps to achieve high-end driving assistance functions.
Even some concept cars have drawn a blueprint for future travel, the steering wheel and gear lever can be stored and hidden, and the car can be changed into a living room and coffee room in seconds.
What is more subversive is that in Musk's definition, if the vehicle can achieve autonomous driving, then the owner can completely lease the vehicle when it is idle to maximize the use of resources and value.
Simply put, it not only generates revenue, but also reduces additional expenses such as parking fees.
For us ordinary users, the essence of large-model artificial intelligence is to redefine the future travel mode and use experience, and even the ownership and use value of the vehicle itself.
But all this needs to be supported by autonomous driving.
Is it to improve safety or increase hidden dangers?
There is no doubt that after ChatGPT gets on the car, it can release many measurement thresholds according to the driving habits of a large number of users and accelerate the development of the autonomous driving industry.
But the reality is that autonomous driving is still out of reach.
For example: Apple Car, which has been developing L5 for many years, has been actively downgraded.
At the same time, in the humanistic concept, people are controllers, and machines can never be replaced, but will a car with deep learning capabilities infringe on the driver's autonomy, resulting in potential safety hazards? How to solve the problem of repeatedly stolen data? There is a contradiction in this.
Car companies obviously have a huge responsibility.
*Image source Motor 1
However, let's take the basic regulations at this stage.
After the vehicle turns on driving assistance, once a traffic accident occurs, the smart car party is responsible, of course, the driver also needs to bear the corresponding damage liability.
In the event of a traffic accident caused by quality or technical failures, the driver shall also be liable first, and then claim compensation from the manufacturer and the seller.
In this way, handing over to artificial intelligence to drive a car seems more unreliable than lending the car out.
Because it has no responsibility ability, once something happens to the accident, the consequences can be made up by themselves.
Even if technical and regulatory restrictions are excluded, if there is a super smart high-end system in the car, do you dare to use it with confidence?
It is precisely because the driver has not been completely liberated in the car, and the large-model artificial intelligence is currently mainly used in more refined intelligent functions.
For example, this time ChatGPT got on the car Mercedes, its voice interaction performance has been significantly improved, such as: it can recommend meals or attractions, when the owner raises complex questions, it can also give more comprehensive answers, and even control smart homes, directly taking the life of people and cars in science fiction movies into reality.
But does this increase the security risk?
Of course, I'm not targeting Mercedes. I'm targeting all the overly intelligent cars these days.
Take navigation, we can stop the car, use the mobile phone to operate, and then the navigation screen is projected to the car machine, and focus on driving again.
But as the depth of voice assistant dialogue increases, it can evoke navigation by voice, but in this way, people need to use more brains, is there a suspicion of distraction?
Also, when we gag with the car machine, why is it not playing with the owner? It's terrifying to think about.
The meaning of voice assistants is to improve the convenience of control, but no one guarantees that it will not increase safety risks.
In addition, now some car machines can also control smart homes, why do they have to operate through the car machine?
The first bundle object of all kinds of home appliance programs is obviously mobile phones, whether they are piled up for the sake of stacking functions... How much does it really mean?
Interior screens are getting bigger and bigger, more and more, and there are fewer and fewer physical buttons.
Is all this progress or regression?
Mazda once believed:
The size of the screen in the car is too large and the functions are too fancy to affect the concentration of driving, so the screen is deliberately changed to simplify the function. However, from the perspective of sales performance, it is obviously not in line with the current purchase needs.
The old fox felt:
In the face of a century of changes in the automotive industry, it is not safe to be too conservative or radical, after all, in the face of all kinds of cool intelligent experiences, richer functions, may require us to pay more energy.
But in the current use scenario, the driver is far from liberated, as Volvo said: the prerequisite for all happiness is safety.
In addition, on the use of smart cars, there are also experts calling:
Driving schools should be reformed and corresponding training exams should be opened, because this type of car is very different from traditional cars, and drivers need to have a learning process.
ChatGPT gets on the bus, is it in a hurry?
Resources:
motor 1、 insideevs.com
Editor: Ze Tian