Navigating the AI Frontier:

How Businesses Can Harness Opportunities and Mitigate Risks of Generative AI

 

Not long ago, artificial intelligence (“AI”) was of little to no concern for anyone other than science-fiction characters.  But in just the last couple of years, the introduction of an ever-increasing number of AI-based language models, like ChatGPT, has taken the world by storm.  Already, businesses around the world—from industry titans all the way to mom-and-pop shops—are harnessing the power of AI to improve their businesses, gain a competitive edge, streamline their practices, and increase their profits.

And as AI technology continues to evolve, it continues to present new opportunities for businesses.  But with great power comes great responsibility, and harnessing this new and powerful technology is no different.  Businesses that use AI must also navigate a minefield of potential pitfalls that could wreak serious harm to the unwary.  Further, governments around the world are abuzz about AI and its societal implications—and how to regulate those that use AI in their business operations.

This article, the first in a series of related articles, will first explain what generative AI is.  Next, it will discuss both the benefits it offers and the risks it poses for businesses.  It will also explore some of the regulations that world governments are already working toward implementing to restrain the use of generative AI in the business world.

What Is Generative Artificial Intelligence?

Generative AI refers to a class of AI algorithms designed to generate content, such as text, images, or even music, that mimics human creativity.  Unlike traditional AI systems that rely on predefined rules or large datasets for decision-making, generative AI models are capable of generating novel outputs by learning patterns from huge sets of data drawn from sources all over the world.  In overly simplistic terms, generative AI makes new stuff by reading a bunch of old stuff.

One notable example of generative AI is the generative pre-trained transformer (“GPT”) series originally developed by OpenAI.  These models, including the famous (or infamous, depending on who you ask) ChatGPT, are trained on vast amounts of text data and can produce coherent and contextually relevant text based on prompts provided to them.  Where search engines like Google work best when presented with short, focused prompts, generative AIs thrive when presented with long, complex instructions.  This technology already has vast potential applications, from content creation and automation to serving as personal assistants and virtual agents.

And we are just at the beginning stages of generative AI technology.  Month by month—and even day by day—developers are hard at work to find new uses for generative AI.  Experts in this field report that even they cannot yet imagine the potential applications for generative AI that will emerge in the future.  Bill Gates, for one notable example, observed that “[t]he development of AI is as fundamental as the creation of the microprocessor, the personal computer, the Internet, and the mobile phone.  It will change the way people work, learn, travel, get health care, and communicate with each other.  Entire industries will reorient around it.  Businesses will distinguish themselves by how well they use it.”[1]

Right now, businesses all the way from industry titans to small startups can tap into the capabilities of generative AI to increase efficiency, foster innovation, and even improve their bottom line.  On the other side of the same coin, however, generative AI poses serious risks that businesses must be aware of before diving headfirst into this powerful technology.  While it is impossible to know just how far AI will go in transforming our lives, and our businesses, there are some immediate steps that businesses can—and must—take to adapt to the ever-changing landscape that generative AI has thrust upon us.

Benefits and Risks of Artificial Intelligence for Businesses

For businesses, the boons offered by generative AI are boundless and ever-expanding.  The following are just a few examples of advantages that generative AI already offers to businesses:

1. Efficiency: Generative AI can automate routine and traditionally time-consuming tasks, such as content generation, document generation, data entry and analysis, and responding to customer service inquiries. All of this can lead to increased productivity and time savings for employees whose time could be better spent on more skilled and specialized tasks.  Further, generative AI can optimize employee workflows by analyzing processes and identifying opportunities for automation and optimization.
2. Innovation: Generative AI can assist in brainstorming and idea generation by analyzing vast amounts of data and generating novel concepts or solutions that humans may not have considered. This can spark creativity and lead to the development of innovative products, services, or strategies.
3. Problem Solving: Generative AI can analyze complex datasets and identify patterns, anomalies, or correlations that mat not be apparent to humans. This data-driven approach to problem solving can uncover new insights, optimize processes, and drive innovation in various areas, such as supply chain management, logistics, or risk analysis.
4. Personalized Service and Advertising: Generative AI can analyze customer data and preferences to create personalized experiences, products, or recommendations. By tailoring offerings to individual needs and preferences, businesses can enhance customer engagement and satisfaction, driving innovation in customer service and marketing.
5. Screening of Prospective Employees: Some businesses have already begun using AI to help screen prospective employees and even provide assistance in making hiring decisions.[2]
6. Improved Resource Allocation: Generative AI can provide data-driven insights and recommendations to support decision-making processes of both employees and executives. By leveraging predictive analytics and machine learning algorithms, businesses can make more informed decisions that maximize returns and minimize risks.

By employing all or some of the uses outlined above—as well as a myriad of others—generative AI can help businesses decrease operational costs and improve profitability.

For some sectors, the use of generative AI has already been revolutionary.  For example, retailers are using generative AI to forecast customer demand, optimize inventory levels, automate customer service, and even detect and prevent fraud.[3]  Manufacturers are increasingly using generative AI to reduce time to market, create digital representations of their facilities, and predict machine failure.[4]  Even healthcare providers are already using generative AI to predict patient outcomes, analyze risks of treatment approaches, and diagnose diseases and disorders.[5]

But with great power comes great responsibility.[6]  Using generative AI poses serious risks that businesses must be aware of if they hope to get the most out of this technology—and avoid pitfalls, including liability, while doing it.

First and foremost, businesses employing generative AI should be aware that most popular generative AI models do nothing to protect any of the information that it or its employees submit to the model.  Stated differently, the prompts you enter into ChatGPT are completely unprotected from subsequent disclosure to third parties.[7]  Rather, generative AI models take user inputs and add them to the datasets that the model already drew from to generate new content.  This is one of the ways that generative AI models continue to learn.  In layman’s terms, generative AI can and will parrot anything you have said to it—to anyone who asks and without regard for whether the information you shared is confidential in nature.

For businesses working with proprietary information, such as trade secrets, this is a serious concern.  Famously, Samsung employees inadvertently leaked “top secret” Samsung source code to ChatGPT while using the model to help improve that source code.[8]  Because of this disclosure, any user of ChatGPT—including competitors of Samsung—may be able to access Samsung’s source code just by asking the right questions.  This cautionary tale shows that employers must implement safeguards and training to prevent similar disclosures of sensitive information.

For businesses that owe duties of confidentiality to their clients, this risk is paramount.  By using generative AI carelessly, lawyers may inadvertently disclose confidential information, and healthcare providers may inadvertently violate HIPAA.  These risks are similarly pronounced for other professionals, such as accountants, financial advisors, and technology professionals.

As a result, any user of any generative AI model must be careful to not disclose any information that he or she would keep confidential under any other circumstances.  When you tell ChatGPT, you effectively tell the whole world.

Beyond carefully using generative AI, businesses may opt to use generative AI models that guarantee the confidentiality of user inputs.  In the legal sphere, for example, LexisNexis’s Lexis+ AI has comprehensive measures in place that assure the confidentiality of its users’ inputs.[9]  Similar, confidentiality-focused generative AI models are regularly emerging for other business sectors.

Generative AI also has sweeping implications for intellectual property rights—both in protecting those rights and in taking care to not infringe others’ rights.  As noted above, generative AI can create content that closely resembles or even duplicates existing copyrighted works.  Naturally, this raises concerns for the owners of those copyrighted works, who may find that generative AI mimicked their work too closely.  On the other hand, users of generative AI models may accidentally infringe someone else’s intellectual property rights when basing their work on a response from a generative AI model.  Of course, infringement of intellectual property poses serious legal consequences for the infringer.  And “ChatGPT did it—not me” is unlikely to be a successful defense.

The rapid advancement of generative AI technologies presents new legal and ethical challenges that the existing framework for intellectual property law may not be able to adequately address.  Accordingly, businesses employing generative AI should stay abreast of developments in this realm and put appropriate safeguards in place to protect their and others’ intellectual property rights.

Below are just a few more examples of how misuse of generative AI can negatively impact businesses:

1. Quality Control: While generative AI can produce content quickly, ensuring the accuracy, reliability, and legality of generative outputs remains a concern. Generative AI models often suffer “AI hallucinations”—instances where the model generates outputs that lack accuracy, coherency, or factual accuracy.  And when AI suffers a hallucination, it still maintains a misleading level of blind confidence in its outp  As a result, businesses implementing AI must still use reasoned judgement, caution, and oversight, and consult with applicable experts, to mitigate the risks of errors or unintended consequences of use of generative AI.
2. Overreliance: Businesses may become overly dependent on generative AI systems for critical tasks or decision-making, leading to complacency or reduced human oversight. Overreliance on AI technologies may increase vulnerability to errors, disruptions, or unforeseen consequences, necessitating a balanced approach to AI integration and human-machine collaboration.
3. Data Security: As with any technology, there are risks of data breaches, cyberattacks, or misuse of generative AI systems. Businesses must implement robust security measures and protocols to safeguard sensitive information.
4. Reputational Damage: Incidents involving misuse, errors, or ethical lapses in AI systems can damage a business’s reputation, erode customer trust, and undermine brand credibility. Businesses must prioritize transparency, accountability, and ethical behavior to maintain trust and confidence among stakeholders and the public. Blindly trusting an AI to come up with the right answer is not enough.
5. Workforce Impacts: The automation of tasks through generative AI may lead to job displacement or changes in workforce dynamics, potentially causing anxiety, resistance, or job dissatisfaction among employees. Businesses should proactively address workforce impacts through reskilling, upskilling, and workforce transition programs to mitigate negative consequences and foster a culture of continuous learning and adaptation.
6. Legal Compliance: As explored below, the use of generative AI may raise legal and regulatory challenges related to intellectual property rights, data privacy, consumer protection, and a long list of other concerns. Businesses must ensure compliance with relevant laws, regulations, and industry standards to mitigate legal risks and avoid potential fines, penalties, and litigation.

Likely Governmental Regulations of Artificial Intelligence

Right now, we are in the Wild West of governmental regulation of AI.  Around the world, burgeoning regulations restricting the use of AI are in their infancy and do not have a long history of enforcement.  In response to the rapid and unprecedented advancements in generative AI technology and growing concerns about its societal impacts, world governments are considering and quickly beginning to implement regulations to govern the development, deployment, and use of generative AI systems.

Focusing on the United States, the Biden Administration published its Blueprint for an AI Bill of Rights in October 2022.[10]  One year later, in October 2023, it also issued an Executive Order governing the use of AI.[11]  Both of these indicate that the Biden Administration’s foremost concerns in regulating AI include ensuring privacy for users of AI, preventing algorithmic discrimination, providing notice to consumers of how AI technologies work, and providing alternatives to consumers who hope to avoid the use of AI.  Like many of the topics discussed in this article, how the Administration will ultimately implement these general policy goals is uncertain.

Based on public discourse and stated intentions of governmental actors, likely areas of regulation that may affect businesses employing generative AI include the following:

1. Data Protection and Privacy: As businesses begin using AI to automate customer service more and more often, governments are likely to implement regulations that place stringent requirements on how businesses must handle and safeguard their customers’ personal information.
2. Disclosures: Relatedly, where the business uses AI to automate customer service, governments may implement regulations requiring those businesses to disclose their use of AI and how they use the technology. Utah, for example, has already passed such a bill.[12]
3. Advertising: Governments are likely to develop regulations that ensure advertisements, especially that tailored for specific consumers or demographics, are not misleading or deceptive.
4. Intellectual Property: Owners of intellectual property are pressing for legislation governing how AI models may use others’ intellectual property and/or requiring companies who own generative AI models to compensate the owners for the AI’s use of their intellectual property. Whether any government will implement these or similar laws is unclear, let alone how those laws would apply in any particular circumstance.
5. Employment and Labor: As noted above, some businesses are using AI to help make employee hiring decisions. While the United States government has not yet adopted any law on this issue, the Biden Administration has already expressed concerns of improper biases in AI—especially in hiring decisions.[13]  As a result, any business using AI in employee management must stay aware of new developments in this area of the law.
6. Industry-Specific Regulations: Certain industries, such as healthcare, finance, and transportation, may have specific regulations governing the use of AI technologies within the industry.

Multinational businesses should also be aware that every country’s approach to regulating AI is different, and they must keep up to date with emerging regulations governing AI in each country in which they operate.  The European Union, for example, has taken an approach to regulating AI that is arguably more comprehensive than regulations implemented and planned so far in the United States.[14]  The Biden Administration has indicated, however, that it hopes to foster collaboration with foreign governments to strive toward international consistency on the regulation of AI.[15]

Even states within the United States are actively working to legislate and regulate generative AI—leaving businesses with the task of keeping pace with new laws even on a state-by-state basis.[16]

In short, the legal landscape surrounding AI is just beginning to take shape, and it will constantly evolve over the coming years.  Right now, we are in the Wild West.  But just as the Earp Brothers and Doc Holliday eventually brought the force of law to the O.K. Corral in Tombstone, modern regulators will reach businesses’ use of generative AI.  Businesses must stay abreast of recent developments in this area of the law if they hope to safely operate using the many benefits that generative AI offers.

Conclusion

This article only begins to scratch the surface of the implications that generative AI has on businesses in the present—not to mention in the future.  Right now, this technology offers powerful uses for businesses that can streamline their operations, improve customer relations, and decrease operational costs.  In the future, the potential applications for businesses are seemingly limitless.

As Bill Gates forecasted, the development of this technology may prove to be just as revolutionary as the semiconductor, the computer, or the Internet.  And generative AI’s implementation may eventually become just as ubiquitous as those technologies.  Businesses that fail to keep the pace may suffer harshly when they find their competition far ahead of them.

But the use of generative AI also poses serious risks that businesses must work to mitigate if they hope to implement it safely and effectively, including ensuring their compliance with a forthcoming, robust, and unpredictable legal landscape that is just beginning to take shape.

In light of the fluid situation created by the proliferation of AI, this article is a first of a series of articles that our Firm will publish to assist businesses in navigating the challenges posed by generative AI and in mitigating the risks it poses.

DISCLAIMER: This article is provided for informational purposes only and does not constitute legal advice.  We strongly advise readers to consult with qualified legal professionals and/or experts on generative AI to obtain advice tailored to their specific needs and circumstances.  Furthermore, the strategies, recommendations, and insights presented in this article are not exhaustive and may not be suitable for every business or individual. 

 

[1] Bill Gates, GatesNotes, The Age of AI Has Begun (Mar. 21, 2023), https://www.gatesnotes.com/The-Age-of-AI-Has-Begun.

[2] Lindsey Wagner, American Bar Association, Artificial Intelligence in the Workplace: The Future Is Now (June 10, 2022), https://www.americanbar.org/groups/labor_law/publications/labor_employment_law_news/spring-2022/ai-in-the-workplace/.

[3] T. Leigh Buehler, American Public University, Artificial Intelligence in Retail and Improving Efficiency (Mar. 4, 2024), https://www.apu.apus.edu/area-of-study/business-and-management/resources/artificial-intelligence-in-retail-and-improving-efficiency/.

[4] Tim Hafke, AlphaSense, Generative AI in Manufacturing (Feb. 2, 2024), https://www.alpha-sense.com/blog/trends/generative-ai-manufacturing/.

[5] Sandeep Reddy, Generative AI in Healthcare: An Implementation Science Informed Translational Path on Application, Integration and Governance, 19 Implementation Science 27 (Mar. 15, 2024), https://implementationscience.biomedcentral.com/articles/10.1186/s13012-024-01357-9.

[6] “Uncle” Benjamin Parker, Spider-Man vs. Wolverine Volume 1 (Feb. 1987).

[7] See, e.g., OpenAI, Privacy Policy (last updated Nov. 14, 2023), https://openai.com/policies/privacy-policy.

[8] Cecily Mauran, Mashable, Whoops, Samsung Workers Accidentally Leaked Trade Secrets via ChatGPT (Apr. 6, 2023), https://mashable.com/article/samsung-chatgpt-leak-details.

[9] LexisNexis, Lexis+ AI Security Information (last visited Apr. 21, 2024), https://www.lexisnexis.com/supp/lawschool/resources/faculty/lexis-ai-security-information.pdf.

[10] The White House, Office of Science and Technology Policy, Blueprint for an AI Bill of Rights (Oct. 2022), https://www.whitehouse.gov/ostp/ai-bill-of-rights/.

[11] Exec. Order No. 14110 (2023), https://www.whitehouse.gov/briefing-room/presidential-actions/2023/10/30/executive-order-on-the-safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence/.

[12] See Utah Code Ann. § 13-2-12 (2023), https://le.utah.gov/~2024/bills/static/SB0149.html.

[13] See The White House, Office of Science and Technology Policy, Blueprint for an AI Bill of Rights (Oct. 2022), https://www.whitehouse.gov/ostp/ai-bill-of-rights/.

[14] Alex Engler, Brookings, The EU and U.S. Diverge on AI Regulation: A Transatlantic Comparison and Steps to Alignment (Apr. 25, 2023), https://www.brookings.edu/articles/the-eu-and-us-diverge-on-ai-regulation-a-transatlantic-comparison-and-steps-to-alignment/.

[15] Exec. Order No. 14110 § 2(h) (2023), https://www.whitehouse.gov/briefing-room/presidential-actions/2023/10/30/executive-order-on-the-safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence/ (“My Administration will engage with international allies and partners in developing a framework to manage AI’s risks, unlock AI’s potential for good, and promote common approaches to shared challenges.”).

[16] See, e.g., Cal. Gov. Code § 11547.5 (2023), https://casetext.com/statute/california-codes/california-government-code/title-2-government-of-the-state-of-california/division-3-executive-department/part-1-state-departments-and-agencies/chapter-56-department-of-technology/section-115475-effective-until-112025-deepfakes (regulating the use of “deepfakes” created by generative AI, i.e., “audio or visual content that has been generated or manipulated by artificial intelligence which would falsely appear to be authentic or truthful and which features depictions of people appearing to say or do things they did not say or do without their consent”).