AI and Marketing Content Generation: Understanding the Risks

July 7, 2023

It doesn’t seem that long ago that for the average person, artificial intelligence (AI) was something of an enigma seen only in sci-fi movies. While AI has been used for years in many applications (smartphone voice assistants, autonomous cars, “you may also like” on shopping sites, social media algorithms, etc.), generative AI has recently been thrust into the spotlight as easy-to-use web-based interfaces have skyrocketed in use and controversy.

AI is a broad term referring to computer systems that perform tasks usually requiring human intelligence, such as learning, reasoning, and understanding natural language. Generative AI is a specific type of AI that focuses on generating content in the form of text, images, audio, and video. While they can be built on different learning techniques, one of the most common is a deep-learning technique called the generative adversarial network, which generates content by training on copious amounts of existing data and mimicking structures and patterns found in that data.

Explosion of Generative AI Tools

Many companies have begun investigating generative AI because it allows for
the quick creation of content with minimal effort. Creating an AI tool is time-consuming and expensive, so most will rely on third-party generative AI solutions, such as those from OpenAI, Stability AI, and Google AI, among others.

ChatGPT is a text generator that has been given much media attention since its release in November. It is an application that uses the generative pre-trained transformer (GPT) language model developed by OpenAI. Google AI created Bard based on Google’s Language Model for Dialogue Applications. To make things more confusing, some applications developed by other companies, such as Jasper and Copy.ai, also use OpenAI’s GPT language model. Soon there will be hundreds of generative AI solutions that will be stand-alone or added to other products, creating integrated sales and marketing solutions. Salesforce just announced its addition of Einstein GPT.

Generative AI is not without controversy. Like many innovative technologies and ideas, how it was intended to be used isn’t always how it will be used. All of the implications and risks are still being worked out, and those who wholeheartedly support it think generative AI will revolutionize sales and marketing. Others are taking a more conservative view, and many companies have banned its use. As generative AI becomes mainstream, companies should look at educating themselves about it and take steps to ensure it is being used responsibly. While there are concerns about cybersecurity—which should be a societal concern as there is a potential for generating nefarious code that can bypass current protections—this article is looking specifically at using generative AI for sales and marketing content creation.

Risks Associated With Generative AI

Although generative AI can be used to craft emails, provide research, write blog posts, consolidate reports, and more, risks associated with it could jeopardize intellectual property, create erroneous information, and negatively impact your company if it is misused.

Intellectual Property

When discussing intellectual property (IP), we are referring to your current intellectual property and the content that generative AI creates. Generative AI is trained on large data sets and information input by users. The tool retains this information to continuously learn and build its knowledge. An issue arises if an employee inputs confidential or proprietary information, for example, asking it to provide a report summary that contains financial data or a quality issue that isn’t public knowledge. That data could be delivered to another user as an output, exposing sensitive information to the public. In addition, the information generated could contain another company’s IP, which may create legal implications with ownership.

Regarding ownership, the question of whether the information can be copyrighted has been brought to light. According to a Legal Sidebar issue by the Congressional Research Service, the Copyright Act protects “original works of authorship.” The U.S. Copyright Office recognizes copyright only in works created by humans and has denied applications to register work created by AI. Lawsuits are pending, so it will be interesting to watch how federal courts will handle the widespread use of AI and copyright.

There is also a question of copyright infringement. Keep in mind that works on the internet are not public domain
as many erroneously believe. However, if an image is downloaded from the public domain, you have permission
to use it. OpenAI acknowledges that the data for training is derived from publicly accessible data “that include copyrighted works.”

Could the output infringe on copyright? According to U.S. case law, if AI had access to copyrighted work and created a substantially similar output, owners may be able to show that the AI-generated content infringed on their copyright.

Inaccurate Information

When using generative AI to create content, users must be skeptical and vigilant about using other sources to validate facts provided as outputs.

Generative AI is trained on available information that may not be accurate and may have bias. In addition, some information may be fabricated. OpenAI states, “ChatGPT will occasionally make up facts or ‘hallucinate’ outputs.” Similarly, Google says, “Bard may give inaccurate or inappropriate information.” Many users have found that ChatGPT will fabricate sources and include fake URLs when asked for citations.

CNET, a tech media site, came under scrutiny when it published 78 AI-generated articles riddled with errors over the course of two months. The company was forced to pause the use of AI for generating content. If inaccurate content is referenced in new content and passed off as fact unintentionally, either by a human or AI, the inaccuracies will be perpetuated until it becomes difficult to decipher fact from fiction.

Impact on Search Engine Optimization

Much discussion is happening on whether generative AI content can negatively impact your search engine results page (SERP) rankings on Google. In February 2023, Google released its new policy on AI-generated content. Google’s ranking system rewards content that demonstrates expertise, experience, authoritativeness, and trustworthiness (E-E-A-T). Google states that “using automation—including AI—to generate content with the primary purpose of manipulating ranking in search results is a violation of our spam policies.” While the policy states that Google will not ban AI-generated content that doesn’t violate Google policies, those seeking to rank in SERPs will have “original, high-quality, people-first content demonstrating qualities E-E-A-T.” Google can predict when AI writes content.

Researchers have been studying the detection of AI content, and generative AI content detectors have been popping up on the internet—although they are not infallible and can predict incorrectly. While generative AI is great at mimicking and following patterns, it has trouble with randomness and variance found in human communication. There is no originality. AI content often contains repetition of words, a lack of natural flow and rhythm, and no personality. You will usually find that it repeats the same ideas worded differently. Its content will likely be written in a straightforward style with little detail. As generative AI continues to be improved upon, there will likely come a time when it is indistinguishable from human-generated material.

Since generative AI follows patterns, lacks originality, and can be repetitive, the blog posts it writes for one company may be similar to those it writes for another if given the same prompt.

Should You Use Generative AI for Sales and Marketing Materials?

We have looked at some of the risks associated with generative AI tools; however, they can offer many benefits, especially for companies that lack the resources to create sales and marketing content. Lower costs, scalability, and increased efficiency are attractive benefits to many. The answer to whether it should be used depends on how risk averse your company is and how your company intends to use it. Generative AI has been progressing at a rapid rate. Educating yourself on the changes (risks and benefits) will put you in the best position to make that decision.