My first attempt at AI research was a disaster. Early on, I discovered what a bad idea this was when it confused Reggie with Ronnie Kray. However, there is a far more serious reason why you should take everything AI says with a large pinch, or bucket, of salt.
That is because AI’s main source of information comes from Wikipedia, the site that you were told at school never to use as it was unreliable. It is not surprising it is so unreliable when anyone can write an article, and with 7 million articles, the company boasts only 600 administrators and 10 arbitrators.
There are also many who question the ethical nature of Wikipedia now, and whether it has political affiliations. Has it become manipulated by organisations and politicians?
The Wikipedia Bias
Wikipedia, founded in 2001, appeared at one point to move from an encyclopaedia to a social justice movement.
Around that time, an editor, sometimes referred to anonymously as “Mr X”, made a list of unreliable and reliable sources that could be used in Wikipedia articles. They were designated red and green. Green sources, considered reliable to be used, included CNN, Al Jazeera, the New York Times and the Post. Red sources are censored heavily and include papers such as the Daily Mail because it is considered too right-wing and conservative.
They are also not consistent with the rules. For example, they say you cannot name people before they are convicted, but Kyle Rittenhouse, who was a minor at the time, was named and shamed in several articles they allowed to be published.
They use two or three radical left sources which make wild quotes about a person. The reader would not necessarily pick up on this, but more concerningly, neither would AI. They take radical snippets and plant the seed of doubt.
Be honest: how often do you check the sources in a Wikipedia article? Neither does AI.
The problem is that a single questionable sentence is enough to make both AI and your readers doubt the entire narrative.
More Than Just Wikipedia
If you think the problem is limited to Wikipedia, think again.
The next biggest contributor of information is Reddit, where Sam Altman is one of the biggest shareholders. The very founder of AI.
Check it out for yourself. Ask ChatGPT to research a topic and see how many Wikipedia and Reddit articles it cites.
So, can you use AI as a research assistant?
If you want to check every source it gives you and every quote, then yes, go for it. Me? I prefer the quicker method of doing reliable research myself.
That is not to say that AI is awful for writers, because next week I will share with you how I do use AI and the time it saves me.
Until next Saturday: Find the time, find the words, find the way.
I’ve since learned relatability is everything - Matt Lillywhite
Totally agree with this. Wikipedia and Reddit are most certainly not reliable sources and AI still needs a human touch.