Are You Trusting All of Your Content to Generative AI?

Don’t get manipulated by the machine

The introduction of newer and more powerful forms of generative AI has understandably had an enormous impact on the field of marketing. Initially heralded as a complete game changer, advancements in AI technology models were anticipated to completely replace the need for human intervention in the creation of written and visual content.  The speed and cost efficiency afforded by machine learning was thought to easily outweigh any potential concerns about the quality of the content or ethical concerns about the way it was potentially sourced.

Now, nearly three years after the introduction of ChatGPT, let’s stand back and see which of those promises have really come to fruition, and assess what’s really working and what’s not.

Pros and cons of Generative AI

There’s no question that generative AI has already had a profound effect on the work of writers and content developers. Here are just a few of the ways it is streamlining the discipline:

• Automating repetitive tasks. Gen AI is capable of automating and accelerating many of the repetitive tasks associated with content creation, such as research, proofreading, formatting, and status documentation.

• Content Generation. As any writer can tell you, the first draft is always the hardest part.  Presented with a mish-mash of random and often contradictory information, it can take a significant amount of time and research to collect one’s thoughts and wrangle them into a cohesive and coherent draft.  Generative AI can help with this process, by generating initial ideas from a human prompt, which at least provides writers with something to react to.  This can speed up the process and enable writers to spend more of their time refining ideas, rather than struggling to find source material.

• Editing. Gen AI is good at offering broad, general writing advice, and can also offer basic checks and makes suggestions on things like grammar, spelling, space count, and punctuation, allowing writers to save the squinting for something more strategic.

• Repurposing content for a variety of channels.  In a discipline like PR, there are many different avenues for getting the message out, including news releases, media pitches delivered via email and IM, blog posts, bylined articles and opinion pieces.  To a certain degree, much of the content gets repeated across these channels.  Gen AI can help automatically wrangle content into a variety of presentable formats.

That said, AI has not eliminated the need for human intervention – nor is it expected to do so anytime soon.  Why?

• Lack of authenticity.  Gen AI excels at recombining existing information to draft summaries, but it lacks genuine creativity. Output is often formulaic and repetitive, lacking the kind of credibility and unique perspective that only a human writer can capture. If you don’t want to sound like a robot, you’ll need to make sure you include human oversight and editing.

• Inability to differentiate.   Gen AI content is pretty easy to spot. Have you ever wondered how, if everyone is relying on the same 3-5 generative AI tools, the content will vary from individual to individual?  The answer is that it won’t. The more the AI models learn from themselves, the more generic it will get. 

• Erroneous or outdated information. Gen AI models are trained on data from a specific point in time, which limits their ability to stay accurate when the information is dynamic.  Inaccuracies can damage a brand’s credibility and even lead to legal issues.

• Diminished SEO performance.  AI-powered search summaries are leading to precipitously sharp declines in traffic for many brands, due to the fact that the summaries are so detailed and complete, the reader doesn’t even need to click-through.  As we operate in the era of clickless search, we need to rethink some of the best practices around keyword use and paid search strategy.  

• Ethical/legal concerns. GenAI has raised numerous questions about what constitutes original and proprietary content. Since the text and images they create are clearly derivative of existing content used to train the models, should the original sources be credited and/or included in any profits associated? This question and others like it will obviously keep lots of IP attorneys busy for decades to come.

The Best of Both Worlds

Without question, AI is here to stay – it may actually be the biggest technological leap we’ve seen since the Internet itself.  It’s also safe to say that it’s still evolving in real-time, and we’ve yet to see all of the many ways it will disrupt and transform our work.  

That said, the challenge will be figuring out how to use it thoughtfully for the best effect, rather than relying on it indiscreetly.  At this point in time, the human element remains essential to ensuring that AI is used ethically and responsibly, and to produce quality content for optimal effect. 

As outlined in Cision’s CASED framework, there are five areas where human input is critical to AI in PR and communications:

1. Context. Human beings need to supply the context to AI models, to ensure that content is relevant – the more specific, the better.

2. Awareness.  As organizations start to implement multiple AI systems, a human being needs to oversee the tools and be aware of what they are doing.

3. Strategy.  Human beings must continue to determine the goals and desired outcomes of AI-produced content.

4. Empathy.  An AI model may give you what it thinks is the “right” output, but is it ethically right? Is it morally right? This is where human alignment becomes vital.

5. Direction.  Humans are still responsible and must retain ultimate authority on what gets produced and distributed.

Looking to humanize your content and take it to the next level?  Let’s talk Let’s talk Helium Communications

###