Small Business Owners Dipping Their Toes in the Generative AI Waters Need To Understand These 5 Risks Before Immersing Themselves in This Powerful Tech
Small business owners are always looking for ways to decrease costs, increase efficiencies, improve customer and client service, and obtain a competitive advantage. Often, these efforts involve leveraging and harnessing the power of new and evolving technology. And no technological advancement in recent memory has offered as much promise, generated as much excitement, and raised as much concern as generative artificial intelligence (AI).
From content generation to customer service to quality control and beyond, this technology can be transformative for small businesses. But it is not without risk. Business owners just starting to dip their toes in the AI waters need to be cognizant of its potential downsides before immersing themselves.
What Is Generative AI?
To understand the risks accompanying generative AI, you must first understand what this technology is, including how it generates and expands the “intelligence” that is its calling card.
Generative AI is based on algorithms that can consume, organize, and process nearly all publicly available information online: articles, videos, books, images, art, technical manuals, music, news, etc. Based on a human prompt, AI synthesizes all the knowledge it has obtained from its “training” to generate new content or perform new tasks. What makes generative AI both impressive and terrifying is how natural sounding and looking its outputs are and how much and how fast it appears to be “learning” and expanding its capabilities and knowledge base.
For example, ChatGPT – the most well-known of all generative AI systems – can produce content on almost any subject that looks and feels as if a knowledgeable person authored the content. For example, if you run a flooring business, you can type in a prompt that asks the program to “write 500 words about the benefits of hardwood flooring.” In a matter of seconds, you will have readable content you could post on your company’s website to advertise the benefits of the service you provide.
Similar programs create content with music, images, and art, including creating new works in the style of other artists or existing works. Enter a prompt such as, “Compose a song in the style of Bruce Springsteen that sings the praises of jelly donuts,” and it will give you a wailing chorus of “The Boss” singing:
Give me a jelly donut, honey, with a sugary glaze,
Gonna take a bite and let it brighten up my days.
Oh, that jelly-filled goodness, it’s a sugary delight.
In the land of donuts, gonna eat ’em all night!
Of course, generative AI has more practical uses for small business owners than faux-Springsteen tunes. AI can generate contracts and invoices, communicate with customers, automate processes, and develop forecasts and insights into customer behavior and market trends, among other benefits. And that is where the following five risks of generative AI may arise for small businesses:
- Quality Control
One of the primary risks small businesses face when using generative AI is the lack of control over the quality and output of generated content. While these systems can produce text, images, and code, they may also produce inaccurate, biased, inappropriate, or off-brand content if not carefully monitored and fine-tuned.
One recent example of putting misplaced trust in AI involved an attorney who submitted a legal brief composed largely by ChatGPT that contained citations to non-existent, wholly made-up cases. Needless to say, the judge was not amused—and the attorney was ordered to pay sanctions as a result. Similarly, small businesses that rely on the accuracy of AI-generated content can inadvertently disseminate information that contains misrepresentations, which can create potential consumer protection problems or other liabilities.
- Unintended Bias
One of the biggest concerns about generative AI systems is that they have been known to generate content that reflects biases present in the data on which they were trained. This poses a significant ethical concern for small businesses. For example, an AI tool may produce an image of a middle-aged white male when asked to generate a picture of a typical CEO. If AI-generated content has the inadvertent effect of perpetuating historical and institutional gender, racial, or other biases, such content has the potential to damage your business’s reputation and image.
- Intellectual Property Infringement
Most businesses know they can’t take others’ content, such as articles or photos, and publish it on their websites or otherwise use such material for their own commercial purposes. If they did, they may face legal liability. That’s why businesses pay for copyrighted stock images rather than copying and pasting photos from the internet. Conversely, no company wants to see others using its copyrighted work.
Since the knowledge base of generative AI includes countless copyrighted works, small businesses that use generative AI for content creation run the risk of unintentional plagiarism or copyright infringement. AI may generate content that closely resembles existing copyrighted material, leading to legal disputes and potential fines. Without careful oversight and manual review, small businesses can inadvertently publish content that violates intellectual property laws. Even if the output images or documents do not closely resemble existing IP, liability may still exist. When an AI tool “learns” how to create content, it may well be creating a “derivative work,” which is also protected under applicable copyright law. The extent to which this process may give rise to legal liability has not yet been widely addressed by courts and is, for the most part, still an unsettled question.
- Overdependence on Technology
Small businesses may become overly dependent on generative AI for critical tasks. While automation can enhance efficiency, relying too heavily on AI can leave a business vulnerable in case of technical issues or system failures. If the AI system malfunctions or is unavailable for any reason, the company may struggle to operate effectively. Further, if a small business does not have robust compliance policies and procedures, such automated errors may be duplicated and multiplied exponentially, creating the risk a company may face class action claims.
In addition, employees may become disengaged or lose critical skills when too many tasks are automated. This can reduce a small business’s ability to adapt and innovate, especially in fast-changing markets.
- Security Vulnerabilities
Importantly, many of the widely available Generative AI systems do not promise confidentiality of input information. For example, ChatGPT saves users’ conversations for training purposes; even when using OpenAI’s incognito mode (which allows users to disable chat history), chat history can be stored on OpenAI’s servers for up to 30 days and can be reviewed by OpenAI’s personnel.
Moreover, generative AI systems, like all software, can be vulnerable to cyberattacks. Newer ventures like OpenAI, or even ones backed by big Tech like Bard (the Google-sponsored version of ChatGPT), may not have the robust cybersecurity measures in place that more established organizations do, making them attractive targets for hackers seeking to exploit AI systems.
Hackers who gain access to a generative AI system can manipulate it to generate harmful or misleading content, disrupt business operations, or steal sensitive information. Protecting AI systems from cyber threats requires constant vigilance and investment in security measures.
There’s a reason so many movies – from “2001: A Space Odyssey” to The Terminator” – have painted a dystopian picture of AI run amok. This technology’s rapid and seemingly limitless advancement and its unpredictable and unintended consequences certainly create fertile ground for such nightmares. While these sci-fi worst-case apocalypses may be outlandish, there are real-world risks inherent with any new and untested technology. Business owners should approach its adoption cautiously to avoid bringing themselves risks and liabilities that could far outweigh the perceived benefits.
If you have any questions or concerns regarding your business’s use of generative AI, please contact Matthew Mitchell at Maddin Hauser.