When ChatGPT decides that it is more efficient running the railroad right through the center of town.
We have all seen the wild headlines:
“I fired my marketing dept and replaced five people with ChatGPT”
“We made up a fake person and she got admitted to an Ivy league School”
Does ChatGPT deserve all the hype?
What does it really mean?
And how can we use it on LinkedIn?
What is ChapGPT?
We have lots of AI in our lives already
- Photo apps store our photos and recognize and archive things like places, faces, and things so that you can search in your photo archive for “car” or “beach”, your photo app will pull them up.
- Netflix archives what you watch, how long you watched it, your likes and dislikes, and comes up with recommended lists tailored to what it perceives are your tastes.
- Siri and Alexa recognize speech, convert it to text and can answer questions
- Automobile Navigation Apps use AI to analyze moving traffic and interpolate user reported data like accidents to predict travel time
Here’s how ChatGPT differs from other apps:
- a tool to search for relevant information,
- refine that information
- and turn that information into content
- ChatGPT is designed to do this by carrying on an ongoing conversation
ChapGPT is free, though there is a premium service launched February 1 called ChatGPTPlus that promises better access and a few other features.
Google is coming out with its competitive offering called Bard. Bard promises more nuanced responses, showing both sides of an argument for example.
As Bard will use whatever info is on the web, it will be more up to date than ChatGPT
And Microsoft now is coming out with “Bing with AI” which – according to Microsoft – will be a much more powerful chat oriented search than ChatGPT itself.
The Good About ChaptGPT (On Face Value)
It’s fast
It’s easy to get started
The results are astounding
But Then You Start Seeing The Limitations
The current version only uses data it was “trained on” through 2021. So don’t expect it to take into account events or discoveries made in the last fourteen months.
Unlike a search engine, it just doesn’t take what it finds and cut and paste it in the results, or just show you a link, but distills what it finds into its own version. In this sense it is original, but the downside is it is only going on what it finds, and it may interpolate info from different sources
It can be wrong and I mean really wrong. Note that both Microsoft and Google when they announced early versions combined with search on Feb 8 both had huge disclaimers about whether results were correct or biased.
And Unanswered Questions
There is talk that ChatGPT will watermark it’s work so that detection will be easy
There are already companies developing apps to detect ChaptGPT or assign probability that AI was used in content generation
At present, anything you create on ChatGPT belongs to you. ChatGPT confers all rights to that info to you. That could change.
If you use the ChatGPT to write a post, are you ethically required to say so in the post?
There are also possible issues with plagiarism. You don’t necessarily know where ChatGPT got its source material.
What Does This Mean For Content On LinkedIn?
Anyone can create content, though the “create” is open to interpretation.
Content will now be cheap. It will potentially flood places like LinkedIn.
Many of the use cases I have seen seem ridiculous:
- For example “how to use chat GPT to write your LinkedIn Profile About Section.” What follows are seven steps involving a lot of cutting and pasting into ChaptGPT and then you are instructed to proofread the result and personalize it. All this for 200 words?
- I have already seen one company whose plugin will comment on other people’s posts for you. The comments are generic, and kinda dumb. (“What a great idea! I can’t wait to read your next post on this!”). I have already seen one person who appears to be using this approach. It was easy to figure out as all the comments looked alike – one sentence, one capability of ChatGPT and just…boring.
- Other use cases are incredibly time consuming: you instruct ChatGPT to write a 300 word post, then get it to write two more versions. Now you – that is you personally – combine the best parts from each of the three to get your final post for publishing. That sounds like way more work to me than just writing a 300 word post.
- Writing posts. So a company may get rid of their writers, but then just have to turn around and hire re-writers and editors. Anything ChatGPT writes must be reviewed and edited.
We Also Don’t Know What LinkedIn Will Do
How will LinkedIn treat content that comes from ChatGPT? An interesting question. If ChatGPT helps generate more content that LinkedIn users want to see, great. But if ChatGPT is used to blindly generate content and swamps our LinkedIn feeds…not so great.
You can bet that Linkedin will figure out how to detect AI generated content or at least make educated guesses. If LinkedIn winds up swamped with mediocre content, you do not want to be one of these companies that LinkedIn detects, as the result will be less distribution of your content.
Your content will have to be outstanding to rise above the crowd.
Even if you want it to write posts, those posts should be carefully proofread. Remember these are going out under you or your company’s name.
Okay, So What Are The Use Cases On LinkedIn?
I see three good uses, and a possible fourth.
1) For content ideas – “write and article giving me ten examples of how ChatGPT can be used on LinkedIn”
2) or content optimization – “give me ten variations on a headline for this article”
3) expanding content or contracting it. ChatGPT can turn a 100 word paragraph into a 200 word paragraph[p, or into a one sentence summary.
4) possibly repurposing old content. I originally thought this was a good one until I did some experimenting with some old content from a customer. The revised ChatGPT versions were not usable or even reworkable. Maybe I need to refine my instructions more effectively, so the jury is still out on whether it’s me or ChatGPT that is the problem here.
But the first three are good use cases, for generating ideas and alternative ways of getting them across. I think ChatGPT’s best use is overcoming writer’s block.
Summary
ChapGPT represents a change. The popular line will be that it democratizes writing. While true, I think that it will democratize mediocre writing. LinkedIn will likely get flooded with mediocre content. That represents an opportunity for people writing good content that people want to read and that people find engaging. So use ChapGPT to help you come up with ideas, but then do the writing yourself. The one thing ChatGPT can’t do is speak with your voice.
Oh and those five people that were laid off back at the start of this article? I heard they got hired to replace the admissions people that an Ivy League School fired for admitting fake people.
