Forum Moderators: not2easy

Message Too Old, No Replies

Automated Insights means automated competition, is that a good thing?

         

JS_Harris

3:24 am on Jul 18, 2015 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



While we, as webmasters, work on our content, writing and copyrighting skills there is an emerging aspect to doing that which we need to consider, soon.

I'm aware of the WW policy on displaying links, and that in some cases an exception is warranted, so I'm going to give a specific example for this article involving a current news event about a well established company that is obviously not mine but is in the news tonight. The news is that eBay has sold their enterprise division but, to me, the story is that a machine is breaking that news. I believe this example can further discussion on the subject.

Specific article on the Raleigh Durham news site: [wral.com...]

I picked this specific article/page because I see it trending in a couple of prominent locations, including on Google news. The footer of the article contains the text "Copyright 2015 by The Associated Press. All rights reserved. This material may not be published, broadcast, rewritten or redistributed" suggesting it was written by or for the associated press and was in fact redistributed on the news site.

More interestingly it also contains the following text within the article itself
Elements of this story were generated by Automated Insights (...) using data from Zacks Investment Research.


Automated insights essentially machine scours the internet to find different tidbits of information about a subject in order to automatically create sentences and paragraphs that, as you can see, can create a cohesive article. If you have no idea what Automated Insights is or what it's doing it's very likely you'd be fooled into thinking a human wrote the article at this point, it's becoming better at fooling people over time.

My question is this: Now that these articles are quite clearly able to land the most prominent spots in places like Google news and any major news outlet that reproduces well trusted associated press articles is it a matter of time before bots write the majority of news articles by dissecting and analyzing parts from different sources to create a new whole? Is it a good thing that sites which post automated articles can avoid being automatically rejected and removed from natural search rankings? Certainly the quality is better than your typical spun content but... it's still a bot.

What does this mean moving forward as a webmaster who takes pride in, and spends time researching first hand, what they publish? Should we all begin managing and improving our own bots so that we can output stories onto hundreds of sites in the time it takes us to write an article manually for one site? How can we tell where search engines draw the line between not enough bot and too much bot? These questions need to be answered now, the example is evidence of that.

iamlost

5:40 pm on Jul 18, 2015 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Given that SEs can not distinguish senseless mashup from sensible content I have no expectations that automated content aggregation will be identified (except by self as mentioned) let alone treated differently.

The real truth of the matter is that the news industry dropped the ball 20 years back and is just now getting back in the game. Others have quietly taken natural language processing (NLP) far beyond Automated Insights. Research and original content assembly will increasingly become automated. The three big hurdles, (1) that content was mostly not digitised, (2) that logic connections could not be made between diverse sources, and (3) that sensible readable synergetic content could not be created by automated systems have been or are being leaped.

It's just that, for those who have found methods that work, the competitive advantage has been in keeping quiet and monetising directly. I expect within a few years some will decide it's time to cash out by selling their services to others.

Leosghost

5:57 pm on Jul 18, 2015 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Reads like it was written by the people who write for ehow..
Glad I'm not in any niches that have readers that would be interested by this kind of shallow fast ersatz writing..
That said..if I can find a way to identify the scraper / bot it uses when it goes to the "cloud"* ( * the word they use on the automated insights website, may eventually mean other peoples websites..the service is called "wordsmith" ) ..it will be blocked with prejudice..

JS_Harris

7:52 pm on Jul 21, 2015 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Great points, ultimately search engines will have to decide just how much automated content they can stomach. A single article might be hard to decipher as being written by a bot but if a site heavily uses the bot it becomes much easier to spot patterns, these won't hide from search for long. Can the associated press be ignored by search for embracing this new technology? Time will tell.

You mentioned news agencies dropping the ball some years ago, I fear they are doing it again right now. Think about it, if they are able to use software to write thousands of articles at very little cost then what's to stop a search engine like Google from doing it themselves and bypassing the news site attempts? Google is already headed in that direction but they don't syndicate the automated content, yet. I'm positive that Google would rather use their bots and algorithms to create news articles than to try and rank automated articles created by others.