Why AI generated articles are not just lazy but often inaccurate - I must admit I had to test AI created content... and it is pish, bland, repetitive, peerless, tasteless, felt like cheating the math exam, oderless drivel with no fun,  and wrong on so many levels... Back to school kids, you are about to get a slap-down... This may hurt: I am unsympathetic!

The world is ever changing, attitudes are changing, and technologies are changing at a rapid pace and faster that at any period in history. If you have been living under a rock for the last couple of years, then you will not know very much about the growth of artificial intelligence (AI) - how AI is making a significant impact on every aspect of our lives, this is including and not limited to the world of journalism and media.

You can't blame humans for trying it on, sneaking round corners and finding shortcuts. This is  especially noticeable in journalism and content driven websites, where as the demand for content keeps increasing, AI-generated articles seem like an attractive solution to fill the void.

It is certainly attractive to the lazier writer, but the convenience of AI-generated articles is very often accompanied by inaccuracies that negatively affect the perception of journalism quality. It spawned my need to type out this article in which we will 'humanly' explore how AI-generated articles, can be inaccurate due to their reliance on searchable content from search engines and how it can result in misinformation, sometimes due to search engine manipulation by companies such as Google, or just plain out-of-date... and humourless, lack a bit of touchy-feely, and reads like one of those YouTube computer generated narration videos.

AI & Searchable Content from Search Engines

The majority of AI-generated articles are created using natural language processing (NLP) and machine learning algorithms that allow the AI to read through vast amounts of searchable data to build a story. Like a game of Chinese Whispers (is that game name even politically correct anymore?), the sourced data is based on a structure made solely of search engine results, which can result in the proliferation of misinformation... we have all seen 'The Science'.

One such example of where AI messes up big time is a company's information and services, which may not always be accurate as it is based on search results, not original data input. For instance, a company may have changed its services or contact information in 2013, but if the search engines or company techies haven’t updated their Search Engine Optimization (SEO) information, the AI-generated article will give out-of-date information which could lead to serious consequences such as financial loss or missed opportunities, even embarrassment.

Inaccurate information can happen when an AI selects the wrong information from a wide range of searchable data. For our humble AI, picking the right pieces of information from large datasets is called “data curation”. An AI's ability to curate that data effectively yet automatically can often be inhibited by the sheer scale and complexity of information that exists online. It can be challenging for any AI to collect authentic information and curate it into a cohesive narrative story that is both engaging and factually accurate... or fun... or with the odd spelling discrepancy. Yup I fail with grammar and spellun like the best of them - it's a personal touch, you are special people. 

You Lie - Search Engine Manipulation by Companies... Really?

The year is 2023, need I say more? Search engine manipulation is another potential cause of misinformation in AI-generated articles. Search engines like Google may engage in practices that prioritize certain content due to their advertisers, or even governmental regulations.... or may not - for journalistic balance.

For example: Let us suppose a company has had legal action taken against them, and it falls under Google’s guidelines to publish accurate and up-to-date information on the case. In that case, they may be forced to push down negative articles and promote positive articles on the company’s behalf, for a few bucks (nudge and wink). Thus, yes we say 'this in human speak - any article created by AI from such search engine data can present an inaccurate picture and misinform the reader... as we have seen, can lead to tragic results as suicides have even been attributed to misinformation, and very recently.

Censorship is another example of how search engine manipulation can create misinformation. Search engines have a considerable amount of control over what content appears online. As mentioned earlier, if Google prioritizes certain content over others, those with opposing views and information may not even show up in search results. By excluding information, search engines are inadvertently (or purposefully for balance) controlling search engine results and manipulating public perception.

The crazy that is the example of the Chinese government’s censorship system is an excellent illustration of the potential dangers and effectiveness of search engine manipulation. In 2019, Google was working on a search engine for the Chinese market that would incorporate censorship requests from the Chinese government to promote “politically correct in China” search results. The search engine, known as Dragonfly, was later scrapped and tgriwn out due to plenty of criticism, and the usual but worth it in this case - protests from human rights organizations and even Google's own employees.

AI-generated articles can be convenient, but they also may lack accuracy due to relying on searchable content from ill informed or malicious search engines, which can contain misinformation due to search engine manipulation... In turn the human manipulation shows through quite clear in a lack of authored content diversity The quality of your journalism or blog may take a massive hit, and the trust of your readers may be negatively impacted...

We all know proof reading is a b*itch, we all still have mistakes for our readers to spot even after proof reading... for some commenters it is a sport: 'Spot The Typo'... AI-generated content may be an essential part of the future of journalism, its true and fair impact can only be realized through continued honest innovation, trusted relevant policies, and heightened awareness to the hazards of misinformation or cheating the system for a quick buck... It is a journalists duty to speak out against any form of dishonesty or fraud... and to be free reporting, honest and unbiased. Even moreso honest enough not to cheat a living... AI does not 'proof read'.