[CentreOfGravity]

Artificial Intelligence. A topic that has garnered quite a lot of interest in the past few years. All this while, we were talking of a future where AI would take over many jobs. That 'future' isn't too far away now...

A few days back, I was discussing with my parents about my dream job, automobile journalism. We were talking about its pros and cons, when they raised an important point - the relevance of automobile journalism with the onset of AI. That was when it struck me; will the current crop of media houses fall prey to AI? Will we have AI generated reviews instead of Team-BHP reviews? If it does indeed take over, how will the system function? Though I feel that AI cannot replace human opinions and feelings about a car, I know that my defense may not be good enough. So, my question to the esteemed members of this forum - Do you think AI will take over automobile journalism along with other jobs? If so, how?

I really hope and pray that automotive journalism is here to stay. We already have platforms like ChatGPT which provide reviews sourced from the internet, will it be long before they are the source? Looking forward to your views and opinions.

[GTO]

Thread moved from the Assembly Line to the Indian Car Scene. Thanks for sharing!

Some thoughts on this topic:

- AI isn't going to give you solid + unbiased + accurate observations that can only come from test-driving the car . AI isn't going to drive a new model and tell you what's hot or what's not. E.g. the Harrier's sensitive steering at speed, only expert reviewers will tell you.

- AI isn't going to engage in an intelligent pricing debate with opinions from real car owners.

- AI is easy to manipulate as - among other things - it is based on existing data. You can bet that car manufacturers will spend crores & crores of rupees to manipulate that data & give you AI answers that favour them. Just like they buy out magazine content & social media today.

- That being said, AI will wipe out the smaller to medium-sized car websites & auto magazines that are already struggling. Many have shut down in recent years and there will be more to come. Reason 1 = AI-generated articles will come to dominate Google / Bing search results (already the case if you search for a comparison between Phone A & Phone B). Reason 2 = many people will start their search on AI. The small-to-midsize automotive magazines & websites that depend on ad revenue will be delivered a body blow.

Net net, there will always be a place for high-quality content in the world. Be it Indian cars & Team-BHP, or international sources such as NYTimes, Bloomberg etc. AI will make things challenging for sure, but the human race will always consume content that is original, intelligent, informative & fun.

I recently read a mind-blowing Bloomberg article about a sneaker Ponzi scheme. AI won't give you this - I happily pay an annual subscription of 20 grand to Bloomberg for quality articles like these. Neither will AI give you a Thar ownership report like this, a motorcycle review like this or an emotionally-charged travelogue like this .

In comparison to such beautiful writeups, AI articles can be JUNK - check out CNET's money-making machine based on AI content.

[Dr.AD]

Originally Posted by CentreOfGravity

Do you think AI will take over automobile journalism along with other jobs? If so, how?

I run a small AI/Data Science company that builds exactly the kind of solutions and business applications of AI you are worried about, and we discuss such questions in our day to day job, with our customers, colleagues, peers, investors and other entrepreneurs too. I am a mathematician, scientist, engineer and finally an entrepreneur and have seen many such new technologies emerge over last 25 years and closely seen how they affect jobs and careers (and seen my own career affected, evolved, and emerged through all this "mess", if I could joking call it so). My answers below come from that perspective.

The short answer to your question: No!

AI can not "take over" automobile journalism as you are thinking of. In fact, based on my understanding (and of course I could be wrong), AI is hardly "taking over" anything and will not take over anything. Most of those are unfounded fears.

AI will for sure revolutionize automobile journalism, or the entire field of journalism, just like it is revolutionizing many fields ! And to a good extent that is already happening. But revolutionizing something is not the same as taking it over or eliminating it, or making it redundant it.

Human intelligence, human creativity, human emotions, especially in fields such as automotive journalism, will always rise above the AI generated text, and the AI generated text will merely be a tool (and a very powerful tool) at the hands of humans to write even better reviews! Of course, a few can try to misuse AI but those will eventually fade out, or will not get any more importance than those "copy-paste" internet news channels that already exist today.

In my own experience, I am seeing that such fears are expressed in all fields - manufacturing, banking & finance, healthcare, corporate governance etc. And we develop AI solutions in all those fields. Eventually what ends up happening, and what will continue to happen at just a lot more rapid rate, is that the humans quickly learn to use these AI tools as their assistants and rise above those to add new value (through their own insights of those jobs and their intuition and creativity) to their own organizations or businesses.

AI + Human is a tremendously powerful combination that will outclass most AI applications in the foreseeable future. Having said that, there are some developments in the field of "Artificial General Intelligence (AGI)", which might be able to reach the same level as human intelligence in some of the job roles. However, that does not exist today (AGI is still a future vision and not a reality today) and moreover, even when AGI comes to reality, it remains to be seen to what level it can match human intelligence in day to day life.

For some jobs which are purely mechanical and repetitive in nature (such as some of the assembly line jobs in manufacturing), those are the most likely to be made redundant by the AI. But those jobs are and were being made redundant by simple "automation" which is and was available in plenty, long before this AI became so popular. So again not much changes there.

The relatively intelligent jobs such as journalism, legal practices, scientific research, medical research etc are least likely to be "taken over" by this generative language models or the models are ChatGPT as you mentioned. These jobs will hugely benefit from the advances in AI, but will not become redundant for humans!

In fact, coming to your specific question, and since you would like to become an automobile journalist as you said, I would say the future looks very exciting. Instead of worrying about AI, think about how you would use this powerful tool as your assistant and how you would use your human creativity to create something next level than what exists today. There are unlimited possibilities and the sky is the limit! I am actually looking forward to AI revolutionizes some of these things.

So in short, yes AI will revolutionize things and I am excited about that and looking forward to the future. And no, I am not worried for human jobs or human roles because I feel confident that we will rise above that and create more beautiful future for ourselves!

Originally Posted by GTO

Neither will AI give you a Thar ownership report like this, a motorcycle review like this or an emotionally-charged travelogue like this .

Very well said and I agree with you 100%, GTO. The kind of reviews you mentioned above can not be created by AI in the foreseeable future. But they can be created much more efficiently, in the future, by the humans using AI as their tool or as their assistant!

[Samurai]

While I am not an AI expert, I did try to become one for 2-3 years before giving up. Now I provide software stack to multiple AI companies, so I do have some exposure to latest trends in enterprise AI.

Sometime ago I answered a similar question, regarding IT jobs. But your question is in generative AI area, would require a different answer.

Recently I read a wonderful farewell message by a business acquaintance, directed at my team. Everyone was very happy to read the glowing tribute. However, my AI antenna started buzzing. The message was too articulate, using words that person seldom uses. Also, the whole message was generic, without any personal anecdotes. Plenty of adjectives, but not one example of why such adjectives were deserved. So, I checked the message at copyleaks, and it reported 94% chances of AI generated content. That's almost 2 sigma, or two standard deviations away from the mean (human content).

Soon most of us will be able to detect AI content either instinctively or by using a AI checking tool. Why should anyone read AI generated content? It is just word salad. Sites using AI generated content will lose viewers in a hurry. Those who write genuine content will continue to thrive.

[SmartCat]

1) These days, automobile journalism mostly equals Youtube reviews of cars & bikes. Manufacturers too give tremendous importance to these reviewers/influencers. Most of them are invited to media launches. Video production quality of the likes of PowerDrift is on a different level. AI cannot help here.

But then, all bets are off if we have an AI powered humanoid in the future that can say "Koi dikkat nahi, bahut comfortable hain" (no problems, it is very comfortable) while jumping over a large road hump.

2) My exposure to AI is limited to ChatGPT/Bard. It is great at analyzing the past news/events but not an event that happened yesterday/today (eg: media launch & review of a new car/bike).

[anjan_c2007]

Maybe AI takes over momentarily, but very soon fakes and doctored views, contentions and absurd concoctions of verdicts for instances on reviews/ road tests will be glaring even to the laymen.

Hence, even if AI overrules real-time common sense and intelligence for a while, it will be very soon shown the backseat, moved to the luggage boot and then thrown out on the road in that sequence, never to return.

[shancz]

Very valuable insights shared from professionals here and this is exactly what AI can't achieve.

From what I understand at present, AI will make the job of searching, comparing and verdicts on reviews or comparos a peace of cake for users and especially folks not well versed with technology by a simple statement "what do you think about this car or this vs that car".

But since the AI cannot feel it cannot share the experience, quirks and issues for people with varying sizes etc.
Since its still in its infancy, what it achieves has to be seen and it looks very promising. But I do feel it has been overhyped and largely misunderstood.

PS : I would love to have the AI writing my posts perfectly in one go, so I don't have to keep editing it for typos

[fhdowntheline]

As long as a human is needed to drive a vehicle a human is needed to review it. AI system reviews wont be able mimic the trouble of heavy set people gauging the comfort of the drivers seat, or short riders trying to tip toe sitting on a motorcycle.

[DCEite]

AI or not, the future of Automotive Journalism is already dead in India.

Look at Youtube, anyone with a Camera phone is an Auto Journo. Heck, these guys, some not even completed +2 and graduation, let alone specialization in Auto Journalism, get regularly invited to Launch events and exclusive media drives, only based on their subscriber base on social media !

Go figure

[Dr.AD]

Originally Posted by Dr.AD

The short answer to your question: No!

I have already given my comments above along with my answer. But based on the subsequent discussion above and some more thought, I would like to add a few more points. Hope that is OK.

1. Automotive journalism is not just writing car reviews. It is much more than that. Further, car reviews themselves are not just text generation. They are much more than that.

Many seem to be thinking that AI can write car reviews, and they are correct to some extent, as far as only the "writing" (or in other words, "text generation" as we call in the AI technical terms) part is concerned.

AI can do "text generation", and to a good extent, it can generate coherent, beautifully written (from linguistic point of view), based on known facts from the past texts, and 'seemingly' logical text. That it can do and will surely do a lot better in future than what it does today!

So only the "text generation" part of Automotive Journalism, or for that matter any journalism will be outsourced to AI and that is how I see it will be an assistant to human. But writing a full review is lot more than pure text generation, such as understanding the true potential and use cases of the car as an owner/driver, understanding how it feels, how it is different than other cars etc, can not be done by AI. That will be filled in by the human expert.

Let me give an example of the above. Only an human expert can drive a new car and come to a conclusion (just an an example) that it handles well and the steering is very precise and sharp, and a joy to use! AI can never understand this on its own. However, once that human expert drives that car and decides that it handles better than all of its competitors, that human expert, in future, may dictate his/her findings to AI and ask the AI to generate a paragraph describing that. That is how I see it may pan out. Personally, I would never do that because I love writing and I will write everything myself. But I do believe a lot of humans under severe time pressure may resort to AI for generating text based on the key inputs given by them as human experts. Whether this is ethical or not is a complex topic and I do not think we know enough to answer that today. But that ethics question will keep coming up again and again and our perceptions of that ethics will evolve with time too.

2. And taking that discussion one level above, automotive journalism is much more than just car reviews. It also involves understanding the trends and new horizons in automotive industry, being able to judge how a particular product is pitched and what are the market demands, being able to spot the emerging trends and new technologies, being able to spot the gaps in either the technologies or a particular vehicle segment, and in general being and expert in the field and being able to comment on the industry as a whole, as and when required, in various contexts. There is no way AI will do most of this. And this is precisely why I said that AI will not take over the role of an automotive journalist, but instead act an a helpful tool to that journalist.

3. In general, lot of current focus on AI, just after the success of ChatGPT and similar "Large Language Models" (LLM: That is the general class of techniques that we are discussing here), is only focusing on the text generation or natural language processing part of AI. AI by itself is much bigger and wider than that, and includes many other fields such as image recognition and processing, speech recognition and synthesis, detection and verification of objects and other patterns from images/videos, anomaly detection, and of course, good old numeric data processing, predictions, time-series analysis etc. So let us not think that LLM are the only AI that matter or that will change the world. Every aspect of AI is changing the world as we speak.

4. One more reason I believe AI will not take over jobs such as journalism is not merely because of limitations of AI. It is not that AI will fail to be good. I believe in a different theory. I believe in the theory of human supremacy and adaptability of humans to no matter what is thrown at them. I believe that no matter how good AI gets, humans will not just give in, and they will find ways to use AI as their tools and do even more interesting, even more challenging, even more creative, and even more value add jobs with the use of AI than what they do today! We as humans will elevate ourselves to next level of thinking and intellectual tasks, while happily letting AI do many tasks which we do today. And that is why AI will never replace humans, but will continue to act as a tool.

5. One analogy that I really like, which is often discussed in AI peer communities, is that of "AI as the new electricity". In the history, when the electricity was invented, there may have been thoughts that it would take over all jobs and can kill humans and can replace humans etc. But as it emerged, humans managed to fully control the electricity, use it for their benefits, and invented new and much more creative jobs for themselves while letting electric machines do most of the mechanical work. That is exactly how I see AI panning out in the future. Just like electricity changed human lives completely and for good, AI will revolutionize our lives to the same degree, but for the good in the end!

So while my answer to this thread remains the same, I felt like adding some more explanation as to why I believe in the answer I gave. Sorry for the long post!

[Chetan_Rao]

AI can wax eloquent about steering dynamics all day, pulling info from manufacturer data, academic publications and what not, but would never be able to explain the feeling behind the half-thrill half-terror grin when you push the nose in a little too eagerly and the tail threatens to spin you around and deposit you in the scenery.

AI will never be able to explain why a 60-odd BHP Figo was more fun around the twisties than cars making twice the horses and costing thrice as much.

I could go on, but my entire point is a machine will never be able to adequately explain a seat-of-the-pants feeling, because it owns neither a butt nor pants. Best it can do is aggregate and repeat what other human owners of said assets have to say.

[ComboCutter]

Most Automotive reviews are scripted with inputs from the OEM on what to say and what not to say (just see one automobile's test and launch reviews and notice they all say the same keywords/highlights). I think AI will be greatly helpful in that if not already. AI can also mimic the tone of the channel/person when trained properly with their previous review scripts.

Humans may be just diminished into copy-editing and presenting the output.

The feel factor: It is still perspective and one reviewer still cannot compare across all automobiles and human populace. AI with sufficient training data can do that eventually. (NCAP and other testing could be improved with AI )

I think gaining data with multiple sensors across multiple automobiles will pave the way for better insightful reviews. (with better EV adoption it can be seen sooner)

It is always easy to dismiss AI in anything enthusiast but it will eventually come and improve the experience.

[Kkumar]

I don’t think AI will takeover auto journalism or for that matter anything involving creativity. Because creativity is human’s forte which AI will never be able to imitate. At the best, AI can be a good assistant to a creative person.

[Sanjunair5]

I am a Marketeer and work in the Tech industry. Until some time ago, I would have wholeheartedly agreed with the views here by many that AI will not be able to this or AI will not be able to do that. I believe it is true to some extent in today's context and maybe into the foreseeable future.

But...(the dreaded 'but'!). Large Language Models have evolved rapidly and are not just generating text based on data it has been trained on. It generates new content and does it intelligently. Imagine if it is trained in engine designs, suspension designs, roads, and track conditions and so on. The model will spit out exactly how a certain car will behave for a certain use case.

And AI evolution is not just restricted to generating text, it can generate imagery, it can generate videos, audio, and possibly everything humans can imagine and even beyond. The pace at which AI has evolved is unimaginable and scary.

So yes, today AI cannot 'feel'. That rules out its ability to bring out automotive emotions that only humans and petroheads feel and that will keep journalists in play (and people in Marketing like me employed!). But it isn't too long until most of the differences between human-expressed content and machine-expressed content are unperceivable!

[Cresterk]

AI can only generate data based on what is already available on the internet. They could probably make a decent used car review a few years after launch but I don't think it's possible for them to have an honest review of a newly launched car because the only data available on the internet would be whatever the manufacturer released. Unless they make some kind of robot that actually goes and inspects and drives the car.