AI-based content generator tools - your opinion?

[CasinosBroker]

Hi,

Would like to ask an SEO related question:

Have you heard or used AI-based content generator tools?

Examples:

• Jarvis
• Rytr
• Copy.ai
• Writesonic

What do you think of them? are these content generator tools any good?

[chaumi]

Tested most (these and others) and yet to find one that would work for me. Some better than others, but definitely not yet good enough to get anywhere near a good quality human writer.

But many work on the basis of starting them off with the right introduction/text or ideas that link the sections/topics...and potentially if you're skilled enough at doing that well you might have better levels of success (maybe I was just not good enough at doing that in a number of cases)

But if you're a non native and hoping that this is the answer to get material in a non primary language I'd say it will help but won't be perfect (and as every day passes, the need to be perfect grows greater).

Even to get a sensible (final) output requires editing effort. Perhaps you could get returns (without any editing) that are good enough for spam link distributions on web 2 and 3 properties or the odd guest post where the publisher is not too bothered about overall quality. Maybe even your own site to push up the quantity as opposed to true quality (and there are some arguably legitimate reasons why you might want to do that).

Bear in mind that what they are doing (not all of them, and to varying degrees) is pulling content off the web that already exists..and feeding it into the result, either partially or spun (although in some cases more intelligently spun than you might have seen in the past).

All of which means...if you take it to a conclusion.....that even if they're good and use becomes widespread...eventually the internet will become even more full of the same content.

Hence the need for content that is differentiated by real quality will become even more pressing, and content at that level will become even more desirable.

[BCGame Affiliates]

I have tested out lots of these and I have found that with the amount of editing that you have to do to bring them up to standard makes them a waste of time.

I think it is going to be some time before these are worth while to use

[gm2891]

You can use one as your writing assistant, but absolutely not like your content creator especially if you don't know well about the product you promote. I assume also that Goggle has already leant to spot the difference between a human written piece of text and AI. Because recently they've found somebody writing news with the help of AI without checking facts. And they certainly don't want any AI written sensitive content on the web like medicine, studies, government-related etc. There are few know details on the internet available on how Google's algo determines the difference.

P.S However, i'm sure that most of these tools will help you beat articles written by retarded human; crazyspeedtech.com/is-it-possible-to-play-casinos-using-bitcoins/

[baldidiot]

TPerhaps you could get returns (without any editing) that are good enough for spam link distributions on web 2 and 3 properties or the odd guest post where the publisher is not too bothered about overall quality.

This is what I'm imagining a lot will be used for. Or sites built to sell links...

It even says on the sites that the content isn't necessarily accurate and needs editing.

Where I have seen this potentially be a good tool is on retail sites for things like product descriptions where you can feed in parameters and key words in mass and have it spit out readable text versions.

[chaumi]

That's an extremely salient point from gm2891....as soon as any complexity is introduced, the result starts to fall apart or become deeply questionable. And the 'writing assistant' perspective is well put.

[chaumi]

Where I have seen this potentially be a good tool is on retail sites for things like product descriptions where you can feed in parameters and key words in mass and have it spit out readable text versions.

Yeah short bursts of text where deeply explanatory coverage is not necessary. In our game, the stuff you could arguably be banging out in minutes anyway straight from your head or experience.

[wonderpunter]

i wonder how many people rank high using these tools? i mean for high keyword difficulty keywords..

[chaumi]

i wonder how many people rank high using these tools? i mean for high keyword difficulty keywords..

None.

The pages that already rank high for lucrative terms are obviously there because they were based on original, well-positioned, well-formed content, age, authority, links, and various other forms of manipulation.....

.....and in some cases downright trickery, as in the cases everyone should know by now where a form of press release is ranking either 1 or top 10 for dozens, if not hundreds, of highly valuable terms with combinations of (seemingly) stolen/duplicate content.

No purely AI generated content will ever be able to compete with that, unless edited well and probably also having all the other forms of manipulation added in a way that's better than what is already existing....unless maybe it's formed around a 'new' topic where there is insignificant competition yet.

[CatalinAMG]

i wonder how many people rank high using these tools? i mean for high keyword difficulty keywords..

I don't think there could be many. Websites ranking high for high difficulty keywords are usually building strong content using their in-house article writers. Sure, there can be exceptions to this rule as the ranking methods are still weird and spammy websites can hold on top positions just because they have a backlink advantage compared to higher-quality (yet less-linked) websites. I strongly believe that sticking to LSI keywords, good phrasing and text adaptation along with a good mix of on/off-page SEO is far better in the long-run planning.

[astra]

From what I've heard before (read on other forums), AI articles work for low competition keywords.
But somehow never read anybody using AI content for web2 properties... wonder if it would work.

[chaumi]

But somehow never read anybody using AI content for web2 properties... wonder if it would work.

There's an argument that says Web 2 is the best place for AI generated content currently, depending on the intent of distribution.....

If it's being published for the links, then it'll be way better that spun rubbish that may have been used in the past.

If it's for actual readers (as a secondary to the main website)...and is being used more to build up a social presence...then it can be unique from what's on the site, but still just about good enough to get away with publishing. If you want readers to adequately interact with it though, in most cases it'll still need editing/improvement/enhancement.

[oldschool]

I used Rytr as wanted to test it. It took forever to create an article as the amount of time editing was enormous. I don't think Rytr is good if you want to have content that can compete for top KW positions. I believe there may be better softwares but currently the best option is write the content yourself. I'm a newbie in SEO so take my advice as such, this is just my experience.

[potentialeight]

I've seen two cases where it was difficult to distinguish AI-written content from human-written content when it had any level of usefulness whatsoever:

1. Information that's very heavy in figures and data
2. Political discussion and trolling

It's amazing for these two spaces, but I don't know of any other ones where it's worth much of anything.

[chaumi]

'Usefulness' as applied here can of course be a term that sits on a broad spectrum, depends heavily on perspective and use cases. But agreed, if it's 'useful' from an intelligent, thought-provoking, and unique or even near-unique point of view that's the intended end result (wink, wink), then the current tools definitely won't cut it.


For data, personally I'd be worried I couldn't actually trust what was being returned, although it may look accurate (as in you wouldn't be able to distinguish it from human-collated data just from looking at it, yes), and in theory it would possibly be accurate (depending on the tool). Comes back to the need for additional editing/checking again.

[john1602]

I really love technologies like AI & IoT. AI based content writing is trending and many people are using it. When i used it first, I thought it would be similar as article rewriting tool but i was amazed by quality content which was generated from it.

[Tommy_odds]

I assume also that Goggle has already leant to spot the difference between a human written piece of text and AI. Because recently they've found somebody writing news with the help of AI without checking facts.

I do not know if Google is already able to spot differences, but it is true the AI content at the moment is really not good compared to human writers. You need to read again the article and make tons of changes. But in few years they will improve drastically, like the videos deep-fake today.

[astra]

Used Rytr and it's not bad at all.
No chance I'd use it on my own site but for sites like quora it's more than enough.

[datstheguy]

The tools, at this stage, are only fine to fight the writer's block but that's as far as it goes. Might also work for content farms who sell backlinks as their main activity to fill up their websites really quickly.

For real content-heavy authority sites, it's just not there yet. Not even close. Mainly, the facts are nearly all the time wrong and made up so the editing needed would simply outweigh the initial time saving.

[Tommy_odds]

Used Rytr and it's not bad at all.
No chance I'd use it on my own site but for sites like quora it's more than enough.

Did not try that one. Will have a look.