Anyone who reads my stuff on any sort of regular basis will know I have been deeply engaged with the potential impact of AI on all of us, since I stumbled across ChatGPT in early December last year. Of particular interest is the apparent potential for efficiency gains, particularly amongst the SME manufacturers I serve.

On one hand, I have been excited by the potential of AI to generate efficiency and expand the operational scale of SME’s. On the other, scared shitless at the potential for bad actors to sneak into our collective pockets and steal everything.

I need to write to think.

It forces me to sort out the stuff swimming around between my ears, as when I can articulate it sufficiently to write about it in some coherent manner, it leads to some level of understanding.

So, here is my list of the good stuff, followed by the bad, as it relates to the core of my business: strategy and marketing, starting with SME’s and the written word.

The good things AI can do for you.

  • Summarising large blocks of copy, even when it seems very messy.
  • Brainstorming; ideas, subject lines, complementary ideas, headlines.
  • Editing and grammar. (I have been using the editing and ‘speak’ functionality of word for years, it is essential to me, and is AI that we now just treat as part of the furniture)
  • Assembling descriptions and fact sheets
  • Looking for logical holes in an argument
  • Repurposing copy from one platform to another
  • Research
  • Outline and first draft.
  • Translation and transcription

The stuff AI is no good at doing.

  • Humour
  • Reflecting current news and events
  • Factual reliability. (Sometimes, it just makes stuff up)
  • Finding a good metaphor
  • Being creative. The great irony with creativity is that AI opens a whole new set of what is possible with visual tools, which can then combine with verbal cloning tools to completely alter apparent reality.
  • Looking ahead
  • Breaking complexity down to ‘first principles’
  • Pouring another glass when faced by a blank page and a deadline.

Then there is all the other stuff AI will do, and evolve to do in the very near future, that is not writing. Graphic design, integrating currently separate digital systems (API’s on powerful steroids) identifying trends and holes in huge masses of data. The impact on medical technology is already profound. When the human genome was first successfully mapped in 2000, the cost of that first success was in the tens of billions of dollars. Now you can send away a sample and have it returned with your genome map for a few dollars overnight.

The key it seems, is to be very good at explaining to the tool what it is you want, in the detail a 5-year-old will understand. As the header cartoon illustrates, being human while driving this stuff will rapidly become the differentiator.

Header credit: GapingVoid.com.