Team-BHP (https://www.team-bhp.com/forum/)

-   Shifting gears (https://www.team-bhp.com/forum/shifting-gears/)

-   -   Drastic increase in computer generated automotive webpages that ruin the user experience (https://www.team-bhp.com/forum/shifting-gears/238963-drastic-increase-computer-generated-automotive-webpages-ruin-user-experience.html)

I was trying to find the difference between Ford Figo and Freestyle and as usual I hit it up on google. Normally the best answer to any google question is within the top 3 or 5 results.
But the top 6 results look like it was written by a computer and not a human. For example: The line "Figo has 1499 cc (Diesel top model) engine, while Freestyle has 1499 cc (Diesel top model) engine" could not have been written by human. Source: (https://www.cardekho.com/compare/for...-freestyle.htm)
The 7th result from outlook was more satisfactory and went something like this - "The Freestyle is essentially a more rugged-looking Figo facelift, just that it was launched a year before the hatchback".
Many such instances makes me wonder, is the internet filled with auto generated articles? If yes, how authentic is it? Like many other skills will we lose our essay writing or comprehension skills sooner than expected? What will be the impact of this on future generations who grow up reading such material?

I have found the same car review published on two different automotive websites. Not a single word had changed in the first few paragraphs.

So I agree with you many people are resorting to such techniques

It wouldn't be computer generated or rather automated. It would more of the content writer not being aware or knowledgable of the subject he is writing about. In these cases, the writer will copy and paste from various sources of the same subject and make content by only changing some words or shuffling them, which is why you may see similar content on two or more websites and they may not make sense grammatically.
Either they are lazy to research and learn and then write it in their own style or just that they have to meet some timelines.

They are auto generated and thats the only way they can do it. They are not going to have some human (that too who knows his cars ) come up write ups for comparison for 1000's of combinations. Thats just impossible.

Most of these sites allow you to compare the Alto, BMW 5 series and the Lambo side by side. So they end up with automated content generators. Some of the better sites would remove these or change the wordings.

Yes, most comparison sites simply use these kind of bot made descriptions. Whether pure bots or lazy copy paste jobs is something that does not matter even. Same thing happens in most popular comparisons like mobiles, laptops, game consoles or other popular electronic gadgets. On the brighter side, that is precisely why forums like TeamBHP thrives even after all the commercial onslaughts.

Tips I found useful to tackle this :

1. If you have time and patience, video search often gives better results.
2. In searches of such topics add the keyword "forum" to spot good discussions like what you find here.
3. Key in the name of the specific forum, magazine or website you know is good in that area in the search string.

Very true, we see a lot of auto generated content these days. We cannot help it but with progress in the AI, the quality of auto generated content too will improve.

We must treat the online content only as a preliminary information. We must experience the products personally, have our own due diligence and have opinions from correct people and credible sources (like T-BHP for cars).

Auto-generated content is not limited to the automobile world. Try searching any news online and you will find the same article on multiple sites (even reputed sites).

The focus is now on quantity and early timeframes rather than quality.

wow, these computer generated webpages are not too bad. I typed 'Hyundai Creta Vs Skoda Kushaq' at https://articlegenerator.org/. I googled the sentences to make sure these are not flicked up some other site. All the sentences seem to be "original".

Results:

Sadly such websites do not realise, they are sacrificing quality for quantity or being there first or cheapest cost per article or whatever.

One of my ex colleagues contacted me recently. He is looking out to upgrade from his Wagon-R and somehow thought I was an expert. Haha.

Anyway we had a brief discussion on what type of car, usage pattern, driving conditions etc and since I know I am no authority, I did what came naturally. I shared the links to several TBHP forums. Official car reviews, ownership reviews, "What car" etc. He promised to sift through the information goldmine over the weekend. In the end, he sent me this.



One more feather in the car for TBHP.

Not only in automobiles, this has happened to not only all products, but services as well. Not only end user products, but tech-products like microprocessors too are compared by third party sites and results can be erroneous. Financial services such as banking and insurance, too, when compared on third party sites are not necessarily trustworthy, at least for being current.

With highly fragmented markets whether it can be in any other ways? Will the quality of such information improve? Remains to be seen. As consumers what can we do: we have to always go to the absolute source of any information we care for accuracy about i.e. the manufacturers own website and further hope that at least they don't err on this!

I work in the digital content business and what people need to realize is just because something has been 'published' online doesn't mean it's reliable anymore.

Literally anyone and everyone with a cell phone can put anything up these days. In the past, the design and layout of the web page was a decent indicator of the degree of professionalism but these days there are several online tools that allow anyone to turn out a professional-looking page. The actual content is another matter altogether.

There are several low-paid jobs advertised in this space for people with no qualifications to just churn out content on a daily basis. 'x number of words on y topic for z rupees'. All that content gets aggregated, reused, auto replicated in online comparisons, auto-generated articles. The idea is no longer to inform or to educate, it's to grab eyeballs, clicks, likes. So a catchy (mis) leading headline with anything published under it is the order of the day.

Sadly a lot of established newspapers and several sections of newspapers and news websites also follow this model.

You HAVE to engage your brain always. Be mildly skeptical of everything you read online. Check the sources, weigh the information against your real world experience and common sense. What are the credentials of the author, are they qualified to speak about this subject. Don't forward anything that you receive just because it conforms to your world-view or opinion. That's how even unreliable or false information gets credibility because the more it's forwarded, the more the chances of 'someone I know well and respect sent it so it must be true.'

Above all "choose your sources wisely."

Thank you for the link. Like all so called 'AI powered' content manipulators, this one also heavily relies on a search engine. It can turn out some manageable content from some source for generic queries, but when you ask for technical queries, it spews crap.

For example I asked for Madras real estate and it pieced together some content from Indian Express and other news media. But when I queried for 'Ultrasonic piezo electric cleaning' it started showing :
New tech to increase boiler efficiency
Recognition of periodontal disease
Dr.Zihun Sen

Content creation is expensive and involves highly qualified people. So many websites plagiarize the content from some other sites. These content cooking apps would never replace a qualified team.

The credibility of most of these results are nowadays lost due to lack of moderation and thirst for ad revenue. Social media is another big monster and better not to go there. If possible it is always best to get expert advice from people directly who are well versed in the matter/subject.

Also searching for a particular matter means that one will be bombarded with ads based on that subject later on. For instance here the search was for the difference between Figo and Freestyle. I'm pretty sure BHP-ian shrinz.vivek would be seeing a lot of Ford ads by now.

Speaking about this problem in general (and not specifically about the link you've shared as an example):

• Indeed this is a huge problem, and a big money-maker too. Thousands of average auto-generated content pieces can usually garner way more traffic than dozens of good hand-written ones. That's the unfortunate truth.

• It's only going to get worse, as artificial intelligence and neural networks get put to use for these tasks. That's going to make it harder and harder for humans to spot it too.

• The problem is that Google laps these pages up, since they've been specifically designed to appease their algorithms.

• Unfortunately I don't see a light at the end of the tunnel, unless Google takes a stronger stance. Either by having some huge breakthrough in their ranking intelligence, or penalizing these pages heavily when caught, lowering their rankings, or relying on human feedback to spot these generated pages.

• Another question is, where does one draw the line? What is an acceptable auto-generated page, and what is not? If it's just side by side data, it's okay, but if it's woven into sentences, it's not??

• Google always talks about their E-A-T rankings (Expertise, Authoritativeness, Trustworthiness). However, unfortunately the tech giant seems to be doing a pretty poor job when it comes to spotting this accurately. Simply put, it's one against a million. The competition that's trying to outsmart Google is too smart, too many in number and moving much faster.

• The same can also be seen on Youtube, where bots put together videos on topics, using stock footage libraries, generated scripts, and text-to-voice. Thankfully Youtube relies on the community for thumbs up/down & comments as ranking factors, so that keeps this junk content from being high up on most search results on Youtube. Google search however does not have anywhere near as strong a community feedback loop.

• Another sad thing to witness is Copy-Paste journalism, where the information is just ingested at one place, and regurgitated at another -- with no value add, critical thought or expert opinion added. I think Google is a little better at spotting this, and also giving the original source more weightage in search results.

• Recently I saw a website that let you copy-paste text into it, and it would re-write it using synonyms, changing words and sentence structures, so that you'd end up with a totally unique (ie. untraceable) content piece. I bet it's a hit with kids trying to do homework!


Seemed a little too good to be true.

It's a direct lift from here : https://www.cardekho.com/india-car-n...ures-27388.htm


Typo lol:

I was a regular shortwave listener to Radio Nederland till the 1980's. The radio station would from time to time repeat their catchphrase:-

"A radio station still run by humans."

This was more than three and a half decades back, when automation was taking over as the 'in thing' in the SW radio domain.

And now, with the profileration of these new auto websites, it is natural that most of the 'Johnny come early's and lately's' are facing identity crises and their one aim perhaps is to make money through advertising after getting good footfalls, with whatever means dubious or not. Action is where the money is !

Only Google can impose strict regulations and if the respective hosting countries have well legislated and enforced IT rules, hopefully we could see many of these perishing.