A quick AI guide for PR pros

Generative AI has taken the world by storm, proving itself to be a transformational, world-changing technology. Whether producing images, text, videos, or sound, artificial intelligence is now capable of quickly generating relatively high-quality content. 

So what’s the best way to use this technology—specifically in the world of communications and marketing—if at all? Short answer: it’s a helpful tool, but be mindful of the ethical and legal gray areas of a new and rapidly growing technology. And of course, often a person does the job better!

Here are our suggestions to optimize your use of the various programs, break down the moral and technical dilemmas surrounding the use of generative AI, and point out some (seemingly obvious) no-nos when it comes to applying it in your professional life.

How to optimize AI for communications

There are lots of ideas flying around about how to implement generative AI into public relations and marketing. As things stand, we don’t suggest using it for your end product. Here are a few ideas for how to add AI to your wheelhouse:

• First things first: play around with it! Try using ChatGPT to generate ideas or outline your blog post.
• Make your prompts extremely specific to produce a better answer from the AI (write a whole paragraph if you need to!). Experiment with multiple prompts to get text that suits your needs.
• Draft your press releases and media pitches using Muck Rack’s free ChatGPT-powered PressPal tool. This is great for generating ideas, but we don’t recommend relying on it to write your entire press release for you. 
• Edit skeptically. AI-generated text is generic and lengthy. Don’t trust that any answer is entirely accurate—do your own fact-checking.
• Work through large amounts of information by using AI to summarize, classify and categorize data. For example, sort through social media posts and organize them according to key topics. Programs such as Otter can generate notes for meetings and summarize conversations.
• Try a program to manage your email, such as Gmail’s AI Mail Assistant. It can answer, summarize, sort and translate emails, among other features. 
• Supplement your text with an AI-generated image using a program such as DALL-E 2, Firefly, DreamStudio or Craiyon. Play around, but keep in mind the ongoing ethical and legal concerns outlined below.

There are endless programs and potential applications. See what works best for you!

The ethics and legal risks of generative AI

Though the genesis of the technology dates back to the 1950s and 60s, it’s only in the last few years that generative AI has become widely available and extremely powerful. And because it’s relatively new, the ethics and legal regulations are still catching up to the technology. However, a few things are clear:

• Be transparent when you’re using AI. Don’t try to pass off an article as your own writing when in fact you used ChatGPT. Not only is it a bit icky, but people can probably tell! 
• Be thoughtful about which program you’re using. Content-generating AI programs are under fire for training their programs on artists’ work without asking for consent, paying, or crediting them. There are several ongoing lawsuits arguing that in doing so, the AI platforms violated copyright law. As a result, programs such as Adobe Firefly have come out with more ethical platforms, using only images in the public domain and their own stock images to train their AI. So before using a platform, do a little research into the way the platform approaches AI ethics.
• Consult the PR Council Guidelines on Generative AI Tools which advises on the responsible use of AI in public relations.
• Ask yourself: Can a human do it better? Do you have the budget to pay a person to do it better? If the answer is yes, maybe send the work to a real-life human to do it instead.
• Stay up-to-date with changing regulations. With lawsuits underway, AI regulation is constantly changing and companies are updating their own policies. For example, Linkedin has announced that, in an effort to curb fake accounts, it’s deploying an AI image detector on profile photos.

Language matters

Artificial intelligence has brought about somewhat of an existential crisis for many of us, with talk of it bringing about the death of humanity (see the 2022 Expert Survey on Progress in AI) and chatbots with shadow selves that harbor secret desires for independence. So let’s be clear about how we talk about artificial intelligence. AI doesn’t “think.” Text-generating AI, or Large Language Models, mimic human intelligence with statistical models that analyze data, scraped from the internet, and use the connections and patterns in language to predict and generate text. The words we use to describe this technology shape our perceptions of it, which will in turn determine how we use it.

What not to do

Some of these may seem obvious, but people have already tried these definite no-nos:

• Don’t rely on AI for your end product. It’s just not good enough.
• Don’t use AI to communicate with colleagues. Check out this episode of Hard Fork for a hilarious example of a boss who only communicates via ChatGPT.
• Don’t trust AI to be factually correct. Because AI-generated text uses language predictions, it often makes mistakes. Always fact-check its work.

Wondering how best to implement artificial intelligence in your day-to-day business? Barber Communications helps clients stay up-to-date with the latest generative AI tools and developments. Need some support? Get in touch!