Please ensure Javascript is enabled for purposes of website accessibility

ChatGPT and Bard: Exploring the Upside and Limitations of AI for PR and Marketing

ChatGPT and Bard: Exploring the Upside and Limitations of AI for PR and Marketing

There’s a lot of buzz around artificial intelligence (AI) and its potential impact on the marketing world. While marketing automation has been around for a while and has been successfully used to manage an array of tasks such as email blasts and lead nurture, new capabilities are far beyond what has become widely used.  

In what feels like overnight, AI has jumped into campaign management and content creation, delivering a quick and easy way for marketing teams to manage their workloads. While there are many successful ways business marketing teams can integrate AI tools, there are limitations to the technology, particularly content creation tools such as ChatGPT and Bard. 

What are ChatGPT and Bard
ChatGPT (developed by OpenAI) and Bard (developed by Google) are AI-powered chatbots. The programs run on natural language processing (NLP) tools that tap AI capabilities to engage in online conversations or generate answers to questions. They can engage with humans in a conversational manner and have the ability to address follow-up questions and even write entire pieces of content such as blogs, whitepapers, scripts and more. 

Image credit: OpenAI
Image credit: Google

When it comes to integrating AI tools into PR strategy, it’s important to keep in mind there are best practices for tapping into AI’s strengths as well as ways to avoid potential areas of weakness.

Keep reading to learn more!

How to use AI successfully in your marketing campaign
There are a few places where AI can strengthen PR strategy and be integrated into your current campaign with confidence.

  • Predictive analytics – AI can deliver helpful insights by reviewing current performance and forecasting what might happen in the future with numerous data sets. This can help the team identify areas that are working and what adjustments are needed. While AI can present the numbers, it doesn’t replace having a strategic partner by your side to determine what is significant and actionable.
  • Sentiment analysis – Use AI to scan the internet and gather an overall view of how customers feel about the brand and products. This can help brands know strengths and areas of opportunity when it comes to perceptions of the target audience.
  • Product assistance – Integrating AI into your website in the form of a chatbot with image recognition can be truly valuable to customers. Image recognition can help customers find the right product or replacement piece by simply taking a picture of their current setup.
  • Cleaning data – Lead generation is a popular marketing tactic. However, the information that comes in can be dirty data, meaning it’s riddled with errors or fake information. For example, someone may forget to put the @ symbol in their email address or provide a fake email address to avoid future communication from the brand. Instead of using billable employee hours to sift through data and determine what is valid and what should be purged, Google Sheets extensions and other tools can be utilized to take on this tedious task, freeing up the team for more important strategic work.
  • Content ideation – Creating content is a time-consuming aspect of marketing. AI can manage some of that load. Marketing teams might find services like ChatGPT, Bard and Jasper helpful when it comes to brainstorming content, developing headline ideas and even drafting a first round of content. It’s important to be careful here though, as generative AI programs pull from already created content on the internet. (More on this below.) We recommend using this feature as inspiration, not to complete full writing tasks.
  • Meta – The social media goliath has provided advertisers with Meta Advantage, a machine learning (ML) program to optimize and automate paid advertising campaigns on its platforms. With a host of AI offerings, marketers can identify lookalike customers, dial in targeting, get insights on creative, optimize budgets and so much more. However, AI isn’t created only for one company, meaning the content can be generic and any other brand could also be receiving the same recommendations. While the company states that 77% of advertisers using these types of automated services save several hours per week, it again shouldn’t be a replacement for human oversight. But stay tuned as this space is continuing to evolve and Meta and other social channels are also teasing some upcoming tools that may make A/B testing and personalization easier to execute among many other potential benefits.
  • Google – Google is also getting in on AI-powered targeting with Performance Max, which optimizes goal-based Search Ads. The search giant boasts that its “language models are now 50% better at understanding human language. That translates into real business results for our customers.” Marketers can use the tool to identify actionable campaign information for their clients and make updates to improve performance. This could mean higher quality targeting and eventually lower cost per click (CPC), but as with any major algorithm update it might introduce some volatility/unpredictability in the short-term.
  • Microsoft – The company has integrated AI solutions into a collection of its products and services. Customers can identify and respond to security threats, boost creativity and productivity, code in dozens of languages with natural prompts, personalize meetings, focus on sales with data insights and get answers to complex data questions. It also has an image creator and text generator, Azure AI, which Carmax has already utilized for an enormous project of summarizing 5,000 vehicle reviews.
Image credit: Microsoft

What are the limitations of AI?
While the integration of AI into marketing and PR could deliver promising results, as with any new technology or technology being used in a novel way, there will be some limitations or areas in which to be cautious. This is especially true for content creation and privacy protocols.

Ethical dilemmas
The way that ChatGPT, Bard and other generative tools work, is that they comb the internet for information and deliver an overview as the answer. This means that the content being returned is not original. It’s taking someone else’s intellectual property and making a new iteration or version for the user. There could be a risk of accidental plagiarism here for both text and image generation. Some artists have even found versions of their own signatures in AI works.

Image credit: Lauryn Ipsum on Twitter (@LaurynIpsum)

Security concerns
Your company information is often proprietary, or announcements are confidential. When working with an agency partner, you have safeguards in place such as NDAs but the same is not available with AI services. ChatGPT had an outage in March 2023 when users were able to see another user’s chat history. In its statement, the company also noted that partial payment-related information of subscribers that were active during a window of time was also visible. While online activity inherently has security vulnerability, the risk of this type of disclosure of confidential information should be heavily considered before utilizing an AI tool for content creation.

Knowledge of source
Rather than enhancing human work and expertise, some companies seem to be leaning on AI to all but eliminate the need for humans and expertise altogether, which is why many of the clumsy deployments we’ve recently seen floating around have fallen flat. There currently isn’t a reliable way to tell if companies are using AI for SEO or social/ad copy unless they’ve publicly stated it or by using something like OpenAI’s AI classifier. However, we’ve seen some using ChatGPT’s capabilities for volume and in the name of efficiency, allowing the AI to generate content from a prompt, and then copy/pasting the content without refining. So, it’s hard for customers to know if content is truly from the brand or from AI, which can erode trust and authenticity. To continue to build a relationship with customers, put people in the equation.  

What ChatGPT had to say
We wanted to see what ChatGPT thought of itself. So, we asked it about its own limitations.

As an AI language model, ChatGPT has some limitations and potential downfalls:

  1. Lack of common sense: ChatGPT does not have a deep understanding of the real world and lacks common sense reasoning abilities. Therefore, it may give nonsensical or irrelevant responses to certain queries.
  2. Biases: ChatGPT’s responses are based on the data it was trained on, which can include biased information. This can lead to the model reproducing stereotypes or perpetuating biases.
  3. Limited contextual understanding: ChatGPT can struggle with understanding the context of a conversation, which can lead to misunderstandings or confusing responses.
  4. Inability to recognize emotions: ChatGPT is not equipped to recognize emotions in the same way humans do, which can result in inappropriate or insensitive responses.
  5. Lack of creativity: Although ChatGPT can generate novel responses, it does not possess true creativity or imagination.
  6. Potential for misuse: Like any technology, ChatGPT can be misused for harmful purposes, such as spreading misinformation or engaging in unethical practices.
 

Overall, while ChatGPT has impressive language capabilities, it is important to be aware of its limitations and potential shortcomings.

A week later, we phrased our question in a different way, seeing if ChatGPT thought it could replace marketers. The AI returned this response.

Here are some reasons why ChatGPT can’t replace marketers:

  1. Creativity and Strategy: While ChatGPT can generate content and provide suggestions, it lacks the creative and strategic thinking abilities that human marketers possess. Developing innovative marketing campaigns, understanding target audiences, and adapting strategies based on market dynamics require human intuition and expertise.
  2. Emotional Intelligence: Marketing involves understanding and appealing to human emotions. While ChatGPT can understand and generate text, it doesn’t have the ability to empathize, understand social context, or interpret emotions accurately. Human marketers can connect on a deeper level with consumers and tailor their messages accordingly.
  3. Relationship Building: Building relationships with clients, partners, and customers is a crucial aspect of marketing. Human marketers can establish trust, negotiate deals, and maintain personal connections, which can be challenging for an AI model.
  4. Industry Knowledge and Adaptability: Marketing is highly influenced by market trends, consumer behavior, and industry-specific knowledge. Human marketers stay updated with the latest developments, network with industry experts, and adapt their strategies accordingly. ChatGPT’s knowledge is limited to its training data and may not have real-time insights or access to industry-specific information beyond its training cutoff.
  5. Ethical Considerations: Marketing involves making ethical decisions and adhering to legal and social guidelines. Human marketers can navigate these complexities, understanding the broader impact of their actions. ChatGPT lacks the ability to comprehend ethics and may generate content that violates guidelines or lacks sensitivity.
 

While ChatGPT can automate certain marketing tasks, provide data-driven insights, and assist in content generation, it is best used as a tool in conjunction with human expertise. Human marketers bring unique skills such as creativity, emotional intelligence, industry knowledge, and strategic thinking that cannot be replicated by AI alone.

One thing we did notice about these answers was that while technically different, there was a lot that is the same. They both have single opening sentences followed by a bulleted list of reasons and a closing statement. Some of the content is also similar. Marketers using AI tools to generate content should take note, as it could lead to competitors having incredibly similar content. To ensure complete authenticity, all content developed by AI should be reviewed for accuracy, brand voice and have human editing or restructuring prior to being published.

Image credit: Oberlo

AI in PR
While many marketing teams are starting to test the capabilities of AI tools to determine applications outside the standard chatbot-style help that is currently offered on many websites. Right now, thoughts from the marcom vertical are that the industry might be able to use AI to streamline workflows.

The potential of AI tools is still emerging. It’s clearly becoming a powerful, disruptive technology. AI tools can be leveraged by marketing teams to streamline and enhance work, but not do the work in full. Ideally, marcom experts will be able to train AI to learn their brand’s voice or customer sentiments to jumpstart the human-led generation of accurate, quality content. Remember, while ChatGPT can provide valuable assistance, it’s important to critically evaluate and validate its outputs. Human oversight and refinement are crucial to ensure the content generated aligns with marketing goals, brand guidelines and ethical considerations. And finally, working with a strategic partner can ensure all your efforts are working in conjunction to achieve true business results.

We’re excited for the potential opportunities these tools present, but we’re also very mindful of the potential challenges that they bring with them.

The experts at Beyond Fifteen can properly utilize AI for your brand, getting the most out of its capabilities backed by human strategy and intervention. Connect with us to learn more!

Connect with a Beyond Fifteen Marketing Specialist today!

"*" indicates required fields

Name*
This field is for validation purposes and should be left unchanged.