Flynnsights // December 8, 2023

The Future Fabric of AI

Weaving together incredible progress and ongoing challenges

Dan Beca

Dan Beca

Chief Experience Officer

What a wild 2023 we’ve had with AI (generative AI, to be more specific). After nearly two decades of limited usefulness, we kicked off the year with an explosion of new tools and tech fueled by OpenAI’s ChatGPT, which gained 100 million users in January of 2023—just two months after launch. Since then, AI has been seemingly everywhere, integrated into thousands of tools.
Screenshot of a graph of AI-related Google search trends
Google Trends data for “AI” (blue) and “ChatGPT” (red)

Fast forward another ten months to November, and we had CEO Sam Altman ousted by the OpenAI board of directorshired by Microsoft, then reinstated as CEO (with a new board of directors) just days later after a threat of mutiny by 90% of OpenAI employees.

Animated GIF of Will Ferrel in Anchorman

Despite all the drama at OpenAI, one thing is for sure: AI is here, and executives are doubling down on their investment in AI technologies throughout every area of the organization.

According to a recent study by Accenture, C-suite execs are planning to increase investments in everything from customer-facing divisions such as Customer Service and Marketing to operational areas such as Supply Chain, Manufacturing, and R&D—as well as administrative areas such as Finance, HR, and Legal.

AI bright spots

The arrival of AI, specifically generative AI and large language learning models (LLMs), has proven to be a huge productivity booster when used well. Users were found to be up to 40% more productive when using AI compared to a control group according to a recent Harvard Business School study.

Tasks that are repetitive, or where the input and output can be easily interpreted by AI, are great places to leverage AI. A few recent examples:

Taking data from a table in a PDF and extracting it to a more usable format, such as Excel, for charting trends over time.
Screenshot of a prompt to convert a PDF to CSV and a link to the AI-generated output

Rapidly ideating on titles for blogs:

Screenshot of a prompt to create blog post title suggestions and the AI-generated response

Beyond what ChatGPT can do on its own, AI is being integrated into other tools to make folks more productive and more capable. 

I am by no means an artist, but with some simple prompting I can generate artwork that would be impossible for me to replicate with my mediocre Photoshop skills.

A square AI-generated image of a Tiger in the jungle
Prompt: create an illustration of a tiger smirking in the jungle.
Screenshot of GitHub Copilot being prompted to generate unit tests
Github copilot demonstration of writing unit tests for a function

Seasoned and junior software developers alike can also use AI for enhanced productivity. 

Github launched a new service called Copilot that allows developers to autocomplete functions by having AI interpret the name of the function. This example may need some editing, but for simple functions or repetitive statements, it can be a huge time saver.

These are just a couple of examples that scratch the surface of how AI can and is making people more productive—and we’re just getting started. 

While great, AI still isn’t immune to GIGO (garbage-In, garbage-out)

AI can certainly do some amazing things, but there are still some pitfalls to watch out for. Remember that awesome illustrated tiger above? Pretty great looking output with a simple prompt. Now here’s an example of creating an image that didn’t go according to plan:
Screenshot of an AI prompt for the generation of a map of the United States of America

Pretty detailed, right? More information than the simple tiger prompt with explicit instructions for how to create a map of the U.S.: colors defined, borders defined, and what “regions” to create based on a list of states. Now here’s the output: 

AI-generated map of the United States of America with mislabeled states

Not exactly as outlined. While AI did a great job understanding that I wanted to see a map of the U.S., it missed hard on the details about the look and feel, as well as the overall information to be represented. It did make up some entertaining state names though (looking at you East Indiania). While the prompt was very detailed, AI was still basing its output on a training set of data inputs, likely including different maps.

Or how about this example of a different kind of “error” in a summary of reviews from Amazon that Marco Arment pointed out.

While AI did its job of summarizing user reviews, it ultimately wasn’t very helpful to the end-user. More context for how customers were divided on the features could have made it much more useful. AI didn’t think to include that as it just summarized what it saw across multiple reviews. 

Social media post highlighting an Amazon AI-generated review summary for a printer
But again, this is just the beginning. As AI continues to evolve, these types of issues should get better over time.

A few things to keep in mind

If you’re using AI to boost your productivity, remember:

  • Trust but verify—when using AI to generate results where you’re looking for information, spend some time validating what AI has created for you. It can still be a productivity booster, but you’ll want to ensure the accuracy is there.
  • AI is great at creating but needs an editor—along the same lines as above, AI can give you a quick boost whether creating content or writing code, but make sure you take the time to edit what’s been created. It’s a good jumping off point, but as we shift to an age of curation and editing, take the time to refine what AI starts you off with.
  • Some models are better than others—At the heart of AI are models trained on data. Knowing some basics about how each model was built can help you use it more effectively. Is it a generalized set of data? Is it a specific domain or topic? Has it been tested for biases? Make sure you understand the data that’s been used to train the model and both the benefits and drawbacks that could come with them.

Looking ahead to 2024

So, what’s in store for the next year? Lots!

According to an Infosys study on Generative AI, companies in the U.S. and Canada are expected to increase their investment in Generative AI by 67% next year. Look for AI to continue to be at the forefront of tech and business news as companies continue to increase their spend.

Along with companies investing more in their own solutions and building AI into their products, OpenAI introduced GPTs, which allow anyone to build a custom version of ChatGPT for a specific purpose. Look for an explosion of custom GPTs in 2024 similar to the boom in new smartphone apps when Apple launched the App Store in 2008.

On top of all this, consumer expectations will continue to grow as people get used to AI and start to expect it as a part of their everyday experience. Everything from chatting with AI to resolve customer service issues to helping them find the best products faster by summarizing user reviews.

So buckle up, we’re just beginning to see how AI will forever change our lives.

Let's Chat

Ready for more peace in your marketing plan? Just give us a call or fill out the form and we’ll get back to you shortly.

Mallory Diamond

Director of Relationship Development

175 Sully’s Trail

Pittsford, NY 14534

Suite 100