When it comes to deploying artificial intelligence (AI) in PR, the debate continues.
Most agree that PR teams managing massive amounts of data should invest in machine-learning technology and AI experts.
But for those working with small data sets–such as analyzing a modest-sized company’s Twitter engagement for 12 months–a manual review of the top 100 posts often will suffice, argues Christopher Penn, co-founder and chief data scientist at Trust Insights.
“You need AI when you have machine-sized problems. You can use human solutions when you have human-sized problems,” Penn writes in a recent e-missive.
He adds that a key to unlocking the intelligence AI provides involves boiling down machine-sized data sets into smaller chunks. Human analysts can then work with those data sets.
[Not everyone agrees with Penn, as you’ll see below.]
AI for PR? Maybe
While we’d all love to magic away our workloads using automation, the concept that PR pros who are not responsible for large data sets do not have to employ AI is good news.
That means you’ll require fewer consultants and external software, a relief for budgets.
On the career end, it’s a sign that doom-and-gloom predictions that automation will soon replace PR and marketing jobs likely are overblown.
Hinda Mitchell, president, Inspire PR Group, agrees.
“While AI bots perhaps can be trained in time to write news releases based on preset determinations, it seems unlikely that AI will replace practitioners anytime soon.”
That’s not for a lack of trying, however. Plenty of firms are touting solutions.
For example, HelloScribe, a software company that offers “AI-powered writing software for communications and public relations professionals,” recently pitched us.
Its software, the company claims, learns “from the context of [written messages] and [provides] relevant suggestions for content.” This speeds up writing, it says.
Without purchasing additional software, however, many PR pros already are leveraging AI more than they realize. For instance, as the Institute for Public Relations (IPR) notes, the software companies that offer media monitoring, website analytics tools and social media listening likely use AI to parse data and draw conclusions.
Use Cases for AI
To determine potential AI needs outside of existing tools, PR pros must have a good working knowledge of AI terminology, says Dr. Timothy Coombs, a professor in the department of communication at Texas A&M University.
“PR people need to understand the basic language of AI just as they need to know the language of business,” Dr. Coombs argues. For instance, he says, PR pros should know the difference between machine learning and simple text classification ( see sidebar)—and why machine learning can provide superior results to computer-based content analysis.
| Three AI Terms PR Pros Should Know
➤ Machine Learning: Branch of AI/computer science focusing on the use of data and algorithms to imitate the way that humans learn, gradually improving its accuracy. (IBM) ➤ Text Classification: Machine learning technique categorizing responses to open-ended questions (such as “how” or “why” questions), as well as organizing and structuring that data. PR use cases: determining urgency of support tickets; organizing brand mentions by sentiment. (MonkeyLearn) ➤ Confusion Matrix: A table used to judge the performance of a classification model (such as text classification). Accurate and inaccurate results of machine analysis can be tracked. Example: calculating whether a prediction that a patient has an infectious disease is correct or incorrect (displaying a “true” or “false” result in binary terms). (Data School) |
This will help PR pros audit the benefits (or lack thereof) of adding higher capability for machine learning and AI at their organizations.
Dr. Coombs’ research focuses on how AI identifies risks and signs of crisis.
“A PR person might miss the warning signs if they are swamped with data,” he says. Yet he aligns with Penn when he says, “But that supposes that the PR people must cope with big data to be effective, [which] may not be the case for many small and medium-size firms.”
Despite the emergence and proliferation of AI firms that take on writing tasks, Dr. Coombs is wary of such claims, arguing that “anything that involves content creation should not be done by AI.” It’s “too complex” for AI at this point.
Still, in small doses, such as the predictive text one might see when typing out a Gmail message, AI can be useful.
“All writers are using some forms of AI in the context of predictive text algorithms, which complete our sentences based on past writing behaviors,” notes Mitchell.
Garrett Jochnau, insights manager at public affairs and digital communications firm Kivvit, is more bullish on AI. It’s needed regardless of company size, he says.
The Counter Argument: AI is Needed Now
For instance, while your company may have a small social media profile—which can be monitored manually, as noted above—AI can help track data from large numbers of consumers, which PR pros then use to make informed choices about messages, product development or brand direction.
“No individual can analyze the behaviors of a million profiles…Breaking the data into more digestible audience or coverage segments is one way to do that, made possible by AI,” Jochnau says.
For market research, a function that may or may not reside on the PR team, “it’s non-practical…to manually segment hundreds of thousands of social profiles based on shared behaviors or interests,” he adds. “AI can do more high-level analysis about overlapping interests [and] influencers. For small- and mid-size firms in particular, AI allows for more advanced insights analysis.”
Regardless of the scope of your social media or public sentiment analysis, AI does not come cheap, Dr. Coombs says.
“Money is the answer to many questions. If you do not need to address large data sets, AI is not a good investment,” he says. “You can do the equivalent work without it by using other computer programs” such as content analysis, “or alerts at a much lower cost.”
Another Option: Watch and Wait
A potential strategy is watching how AI is evolving and waiting until it’s right for your needs. As uses of AI evolve, the technology should improve, become more widely available and drop in price. Since machine learning becomes more accurate by analyzing more data, “that means later versions of AI will have greater accuracy. If you can wait, or lack of funding requires you wait for AI, that is a viable strategy,” he says.
Yet watching and waiting are unlikely to yield a perfect solution. “AI is not a magic cure-all for every organization or all PR tasks,” Dr. Coombs says. He advises PR pros consider “their needs, resources and possible effect on the organization when deciding on whether, and how, to use AI.”
Contact: [email protected]