Day 16 – AI

Prompt 16/20

Blog: Tell us your hopes (or fears) for the future of AI in WordPress. Post your response on a WordPress website and link it in the comments.

Hopes.

I think this Tweet and video from Joe Hoyle sums it up pretty well.

The conversational nature of it is something I think has massive potential. I am a Web Developer and as such I know the general lexicon – row, column, font sizes, tables, hex colors… which might seem natural to us but is not to a wider variety of people. One comment I hear over and over again from clients (and designers) is that the WordPress admin is overwhelming, there are too many options, too many ways to break things (unintentionally and intentionally). This might help a lot.

Using something like AI makes it (potentially) more accessible to a wider variety of people, as great as we make any UI – adding another column to a table is going to be a different process from system to system. The words ‘Add another column’ will not be. Obviouvsly we know AI can ‘hallucinate’ and so this is not guaranteed. More on that next.

Fears

Well, many. AI or for many contexts LLM is not magic. They appear to be but they are not, however the expectation, even if known (myself included) is such that they are. As such we might over trust it to produce the right results even if they are not. The need to check these is essential.

Auto generated fluff. And there will be LOTS. If as above we can generate things more easily then this will be for good and bad and maybe both. If we can generate an essay from a simple prompt then we will, and a LOT. SEO spam, filter content, marketing gateway pages and much more Ray-Ban style comment spam than you can shake a Midjourney generated stick at.

This will undoubtedly be overwhelming, and also mean we will question the nature of truth, if anything can be AI and ‘made up’, what is ‘real’.

People will develop AI detectors which will also be AI powered and will be wrong and right in equal measure. See how it’s affecting education already. Maybe I’ve put this blog post through ChatGPT, maybe I asked it to fix the spelling and grammar, maybe I asked to to generate the entire thing, maybe my prompt was super vague…or super detailed? All are AI generated but some are orders of magnitude different to the other. Is it any less valid or true if I just want to fix my spelling. What if I was dyslexic? What if my English was poor. There are so many nuances it’s almost impossible to say AI is bad, AI is good, AI is false. It just is. All neatly bringing me around to the first post on this blog in 2004.

Or another way of putting it.

“All of this has happened before. All of this will happen again.” – Number Six

View All Posts