My opinion here is from a complete non-techie perspective as I know nothing about the workings of AI at a technical level apart from a broad sketch from a laymans perspective.
In my area of work, financial markets, over the last 2 decades the kind of the ease with which one could analyse the markets has grown exponentially. Earlier a few softwares could give one access to various financial metrics of different companies in a concise and easy to read manner. Now we are at a place where the system itself can compile, analyse, run simulations and then optimise the strategy and throw out an investment/trading idea. And I believe we are still at the beginning of this evolution. To do the same job earlier a Research house would have employed many juinior analysts then a few senior analysts and then a manager at the top who would ultimately take the call. With the current capabilities one can easily remove lot of the junior level employees.
Same is the case with customer care industry where many of us are already chatting with a bot. Even now these bots are pretty dumb and one can easily recognise if they are chatting with a bot or human. But with the pace at which the natural language models are evolving it may not be a stretch to say that many of those jobs will be gone. Of course jobs will probably shift to tech side, but just like automation in manufacturing sector, the number of new jobs created may not be equal to number of jobs made redundant. Also it is my opinion that while the new tech jobs create/bring in more value in ecomic terms, the number of jobs are normally very less compared to the jobs that were made redundant, especially in areas like manufacturing.
Now obviously all these examples are from non-creative aspects or applications. But I have read about few companies using just AI to create video content for their marmeting purposes. And not the animated version or such, a complete impersonation of a real human, which, even the said individual was very impressed with.
Guardian article of the same.
Many of us might have also seen AI generated images of real and imaginary people. As of now this is restricted to fun part, but when serious actors and big players move in it may change entirely. And as a few pointed out, majority of content in the creative sectors, especially the text part is very generic and just a rehash of some event with different words or style. AI can actually do a far better job than many with it's ability to be better and grammar(I myself suck at grammar).
Coming to the primary part, journalism in general and automotive journalism in particular, I can envision a hypothetical (I believe this is possible) scenario where an AI model can provide far better environment for a customer trying to read a review of a vehicle than majority, if not all, journalists. I am of the strong opinion that majority, including people like me, aren't the typical T-bhpian who is very adept at the technical aspects of a car. Also many a times the opinions can be vague or subjective which may not properly transmit the idea that the reviewer wants to convey.
Take the below hypothetical example/situation.
Modern cars come in with lots of tech with many features requiring some sort of connectivity. In near future a car connected to internet or a server of the manufacturer may be default. I magine the amount of performance data that can be gathered from every single car anonymously and put that data to compare the performance of different models across n number of situations/conditions. Leave alone real world data, a portal can have a device plugged in and review all the various models and collect the data across a set of similar conditions on a track designed to reflect the various driving scenarios. One can probably completely automate the driving itself if the track is designed in such a way that just with a device plugged in one can automatically drive it with a software. With the amount of data collected they can now objectively tell which car performs how in which scenarios. The same can be put in simple sentences given the evolution of language models.
One can correctly say there is more to driving/review than just numbers and analysis. But reviews can be highly subjective. Two different people driving the same car can come out with opinions ranging from being same to diametrically opposite. Now imagine an interactive AI persona that can interact with every customer and give them the analysis in simple language based on their own requirements. Say, I want to know the performance of a car at speeds between 80 and 100 specifically. The AI model can analyse the data across the various car tests and put out the results and describe the performance in simple to understand language along with providig data to those who want to dig deep.
Does that mean automotive journalism will be dead. Not quite. Just like in an age of precision engineering a hand made watch still commands some premium and is a niche segment. Similarly there are companious in ultra luxury segment which market hand made vehicle as a niche. So would be an article written by a human.