ChatGPT, Gemini (formerly Bard), and many other machine-learning applications are being used as chatbots, virtual assistants, or to answer questions on your college exam, but how are marketers using these kinds of tools, and to what end can they be used to amplify and enhance your online marketing efforts. For the past few years, services like Jasper.ai and other AI copywriters have been available, and their main target audience is marketers, such as writing “high converting copy” for your ads, web pages, and social media posts. For the most part, these tools are fairly innocuous, apart from the ones that Twitter trained, but that leaves me with one big question since people make these tools, and the answers are derived from other people’s work:
Should online marketers use AI to streamline processes?
In short, yes, you should use any method to help automate and optimize your strategies to amplify your business online, such as using AI-generated content for your SEO efforts, but there are a few caveats:
- When and where was the data set derived for the AI?
- Are you modifying or transforming the content to make it your own?
To better understand the ethics behind AI-generated content, you first have to understand how these systems work.
The accuracy of AI-based content is based on its data set.
ChatGPT, in its current form as of February 2023, is utilizing a snapshot of data up to September 2021, so if you are asking about anything developed past that, it could return inaccurate results. Additionally, these results are based on the data from which it derived its initial learnings. The internet is a vast, unregulated space with lots of misinformation, or just fanfiction, out there that can mislead or downright lie to readers, so similar to when one does a research paper, make sure the cited sources are trustworthy and accurate.
Are you changing the content to be your own?
Content derived or written from an AI can be pristine and accurate the moment it is output, but you also need to ensure you aren’t copying someone else’s work. To ensure AI-generated content is both unique and yours, provide highly specific prompts, curate the training data to reflect your brand voice, and rigorously edit the output for accuracy, originality, and brand alignment. This way, AI becomes a tool to enhance your creative vision, not replace it.
The earned experience from years of marketing should not be discounted with the introduction of AI. As consultants, we should augment our campaigns by leveraging AI for data analysis, leaving more bandwidth to make meaningful insights and data-backed decisions.
— John Gibbings, Director of Content Relevance
How do ChatGPT and Other AI-Based Chatbots Work?
What exactly makes up AI-based chatbots, for this example, I will primarily focus on how ChatGPT works. The backbone of ChatGPT is Natural language processing (NLP) which refers to a sub-field of artificial intelligence that enables computers to interpret, understand, and generate language similar to how we (as humans) do. Natural language processing (NLP) is the cross-section of linguistics and computer science, and popular examples of NLP are autocorrect or autosuggest. The NLP model will segment and reduce a search phrase or question down to its base form, such as removing suffixes, stop words, and dictionary forms of words, then each reduced term is applied a specific tag such as noun, verb, adverb, etc. Then proper nouns or named entities will be tagged appropriately in this tagging process, and this helps a computer model understand the meaning of the words and sentences. This work helps encode a specific phrase into data that a computer can interpret, 1s and 0s, or binary.
Transformers are a neural network that excels at understanding long-range relationships in sequences, like text. They achieve this through attention, which allows the model to focus on specific parts of the input sequence that are most relevant to the current task. This makes them particularly effective in NLP tasks like machine translation and text summarization. In generative search models, transformers can analyze a user’s query and identify the most relevant information from a vast text corpus. They can then use this information to summarize the search results concisely and informatively. This makes them a powerful tool for helping users find the information they need quickly and easily.
Large language models (LLMs) like ChatGPT and Gemini are trained on massive datasets of text and code. These datasets can include:
- Books: This exposes the model to various writing styles and vocabulary.
- Articles: News articles, research papers, and other forms of factual writing help the model understand the real world and current events.
- Code: Code teaches the model about programming languages and how computers work.
- Web crawl data: This vast collection of text and code scraped from the internet. It can include everything from social media posts to product descriptions.
What Kind of AI-Based Platforms Are Available?
There are dozens of ChatGPT alternatives that function similarly; here’s a quick list of the ones developed by big names in the industry as well as a few other notable inclusions:
- META – Blenderbot 3
- Google – Gemini (formerly Bard), LaMDA, Socratic
- Microsoft – Chat-GPT, DialoGPT
- NVIDIA – Megatron-Turing NLG (in association with Microsoft)
- Jasper.AI
- ChatSonic
Many of these chatbots are integrated into other systems, and others are marketing purely for … marketers. This list is by no means a ranking or an expansive list of ALL of the chatbots out there. Still, it is a good place to start in learning how large, multi-national corporations are developing these platforms and how other entrepreneurs are filling in the gaps with their spin on these chatbots and how they can be properly utilized.
How Can You Use LLMs to Augment Online Marketing Initiatives?
LLMs like ChatGPT and Gemini can be powerful tools to supercharge your online marketing across various channels like SEO, paid search, and social media.
Generative AI & SEO (Search Engine Optimization):
- Content creation & ideation: Generate drafts and outlines for blog posts, product descriptions, and landing pages, saving time and resources in the initial research phase. Creating content with AI, brainstorming content ideas, and identifying relevant keywords to target are also good options for generative AI.
- Keyword optimization: Analyze existing content and suggest improvements for readability, keyword integration, and overall SEO strength.
- Competitive research: Analyze competitor content and identify ranking opportunities using LLMs’ ability to process large amounts of text data.
ChatGPT and other generative AI platforms prove themselves to be a boon in SEO, cutting down greatly on menial tasks while also proving itself capable of more demanding projects, such as identifying potential persona groups and guiding content creation.
— Alex Dahms, Digital Consultant
Generative AI & Paid Search (Pay-Per-Click Advertising):
- Ad copy generation: LLMs can create compelling ad copy variations for different audience segments and test their effectiveness.
- Keyword research: Similar to SEO, LLMs can help identify relevant keywords with high search volume and low competition for your paid search campaigns.
- Landing page optimization: LLMs can analyze landing page performance and suggest improvements to conversion rates.
Some of the more interesting ways our Paid Ads team is experimenting with ChatGPT centers on audience targeting, as well as tailoring ads based on user personas and perspectives. For example, the tool is especially useful for identifying potential pain points for users and then generating content/creative ideas that can be applied to assets, encouraging stronger engagement.
— Nina Martinez, Director of Advertising
Generative AI & Social Media Marketing:
- Content ideation: AI can help generate engaging social media post ideas that resonate with your target audience.
- Content creation: LLMs can help craft captions, tweets, and other social media content in various tones and styles.
- Community management: Use LLMs to personalize responses to comments and messages, improving customer engagement.
LMM’s can help speed up social media content creation, however, it’s important to always keep the content relatable and human first.
— Angela Cowo, Digital Consultant
Human oversight is crucial; while LLMs can generate content, human editors should always review and refine the output for accuracy, brand voice, and correctness. The quality of the content generated by LLMs depends on the data quality they are trained on. Ensure the data used is relevant to your target audience and free from biases. LLMs are tools, not a magic bullet. Develop a clear marketing strategy and use LLMs to enhance its execution.
Is it Ethical to Use ChatGPT and Other Chatbots for SEO & Online Marketing?
There is a bit of nuance to the ethics of ChatGPT and AI similar to it, as it raises several more questions in regards to where the learning and information were derived and the manual, human effort that went into the chatbot itself, how often is the dataset updated, and other aspects. Crediting sources in AI-generated content isn’t about traditional citations but transparency. Disclose AI use and acknowledge the general areas of knowledge it draws from (research papers, news articles). Focus on the final product’s originality through your editing and unique perspective. As AI evolves, so will ethical guidelines, stay informed for responsible use.
A discussion in the world of artists right now is based on these AI-art programs that you put in a word or phrase, and it will generate art or one where you input a picture, and it outputs a piece of art that is similar to the one you put in, such as portraits. Art, in and of itself, is inspired by many things, and artists can inspire and interpolate other artists’ styles to make it their own. Can the same be said for content derived from a data set such as ChatGPT, where the original creators of the content, users across the entirety of the Internet, haven’t explicitly permitted their content to be used in this way? Even further, Google does this with featured snippets, Alexa does it with answers through its in-home platforms, and Bing does this through its featured snippets, so are all of these taking other people’s intellectual property and posing it as their own? The big difference between AI-generated art or content and Google/Bing is that there is attribution on the latter, as when featured snippets are placed in search results, a link to the source is placed right there. It is even highlighted on the page if you click that link from the SERPs. Inspiration is how we innovate new things, but we want to ensure proper attribution is given to those who inspire us.
There is no straight answer of the ethical implications of utilizing generative AI models for online marketing initiatives, including content creation, image generation, and much more. Still, it is riding a fine line between fair use of knowledge publicly available on the internet and infringing upon intellectual property. Whether you use AI or not, being open and honest with the stakeholders is the most important issue when utilizing these tools, and not falling victim to pitfalls like the New York lawyers that were sanctioned for using fake ChatGPT cases in a legal brief.