Blog

subscribe to RSS feeds


« back to all blogs

5 Musings About AI & ChatGPT


Hi!

Human writer David Telisman here. In the months since I’ve written about ChatGPT, the world of AI has continued to explode, and I still haven’t been replaced by a chatbot. 

Just in case, I am planning my next career move, and I’ve whittled it down to:
  • Shepherd 
  • Scapula Model 
  • Handyman Who Can’t Fix Most Things 
Or I can stop questioning my fungibility, sit with the uncertainty, and adapt. Sounds like a plan and so does sharing my five latest musings about AI and ChatGPT.

But a scapula model…

1. AI Is Nothing New


It’s funny that people are now saying, “Wow, AI is all the rage,” and “Have you heard about AI?”  

The truth is that AI was conceived over 70 years ago when British polymath Alan Turing authored a paper called Computing Machinery and Intelligence. But Unlucky Alan (sounds like a Trump nickname!) didn’t have the funding to apply his research, because the cost of leasing a computer in the 1950s was similar to today’s life insurance payouts. 

Five years later, Allen Newell picked up where Unlucky Alan left off, initializing the proof-of-concept and securing funding by the RAND Corporation. In this chapter of Allen on Alan crime, Allen emerged the victor, creating the first AI program called Logic Theorist. 

Fast forward to today, and the top applications of AI are in:
  • Healthcare
  • E-commerce
  • Robotics
  • Finance 
  • Facial recognition
  • Marketing
  • Social media 

2. But ChatGPT Is Very Much New


AI runs in the background of most things we do from shopping online and TikToking to banking and streaming content. 

What’s new is that we can use ChatGPT and get instant feedback. AI is no longer only this nebulous, algorithmic activity that’s doing its thing in the ether that we’re not steering. Now we have direct access and control, and it’s turnkey and lightning-fast. 

ChatGPT is on version 4.0, and it now features plugins that integrate apps like Expedia, Kayak, and Instacart, among others.

3. Chatbots Are Only As Good as You Command Them


Rubber-stamping an article from ChatGPT and publishing it is like running into the surf and ignoring all the shark warning signs. It can come back to bite you. 

Using ChatGPT efficiently requires skill. You need to be clear and precise, use follow-up prompts, try different phrasing, cross-check information, and apply iterative refinement. 

Otherwise it sounds very AI-ish, and you know, you can get sued for libel and stuff. 

4. ChatGPT Has Lots of Limitations 


ChatGPT is almost 8-months-old, and while it’s poor form to pick on an infant, I should emphasize its limitations:
  • Its knowledge base runs only through 2021
  • You can’t count on its accuracy
  • Links and footnotes often redirect to out-of-service webpages and incorrect sources
Last time I mentioned that the first prompt I gave ChatGPT was, “Tell me why ChatGPT can’t replace human writers,” and I was pleased and reassured by its response. I still am, because it added reasons like:
  • Lack of originality and creativity
  • Absence of personal experience and empathy
  • Limited by training data
  • Lack of critical thinking
  • Unintended bias
However, it coyly closes with, “The true art of writing is a uniquely human ability that AI, as we know it today, can't replicate.” 

5. OpenAI Will Beat the Rap


OpenAI, the parent company of ChatGPT, is currently under investigation by the Federal Trade Commission (FTC). 

According to tech blog Fagen Wasanni:

“The FTC is demanding that OpenAI provide records relating to the company’s evaluation of the risks associated with its AI models, as well as complaints filed by individuals who may have been defamed or disparaged by the chatbot’s output. The agency is particularly interested in claims of false, misleading, disparaging, or harmful statements made by ChatGPT.

The investigation also stems from an incident in which the chatbot displayed private data from certain users to others. OpenAI attributed this security failing to a malfunction in the user request queueing process.”

OpenAI leaders say they will cooperate with regulatory efforts, but haven’t we seen this movie before with tech giants like Facebook and Twitter? After Congressional hearings, those platforms are still fostering hate speech and the spread of disinformation. OpenAI’s valuation as of May was $27-$29 billion. 

Good luck, FTC.

Embrace It Or Be Left Behind


To be an AI holdout is to stick your head in the sand, and to react with hostility is a fool’s errand. 

I am heartened that when I speak about AI to audiences, they are leery and standoffish about its meteoric rise, and they remain advocates of my craft. It means a lot. But you’re doing yourself a disservice if you don’t give these tools a whirl. 

OpenAI keeps adding to its suite of services, now offering image generators and the ability to integrate OpenAI models into your application or business. There are so many more generative AI applications on the scene that create audio, video, and branding content, as well as automation. 

I can’t predict exactly where AI goes from here, but I’m confident that it will continue to evolve and innovate. As writers, we have to do our part to keep up or even get in front of it.

The prevailing concern for us is that we’ll need to fight and compete against AI, but I think it will be the other way around.  

About the Author, David Telisman








I am a Writer and Content Creator, and I work with businesses to inspire their customers to buy from them. I believe that my clients deserve to feel proud of how their content marketing looks and what it says, and I deliver by providing expert copywriting and marketing solutions.

Subscribe to our blog and YouTube channel, and follow us on Facebook and LinkedIn.

by

 

Blog Articles

Blog Archives

Categories