Balancing Innovation with Security: The Ethical Challenges of Generative AI

By Dr. Briant Becote, Cybersecurity Professor at UAT

• Generative AI’s potential comes with risks like bias, data breaches, and reputational damage, requiring strong ethical oversight.
• Cross-functional ethics committees can help prevent unintended outcomes from unregulated AI use.
• AI errors, like those from Avianca and CNET, highlight the need for human oversight in AI deployments.
• Protecting data is critical in AI projects, as breaches like T-Mobile’s reveal the costs of poor security.
• Working with regulators and ensuring transparency can help organizations navigate AI standards and build public trust.

Artificial intelligence achieves incredible speed and ease in solving complex challenges but also raises profound ethical and security questions. As a Naval Officer (recently retired), I directly witnessed the expansion and evolution of artificial intelligence as a tool for warfighting. I provided international guidance regarding autonomous drones operating on land, sea, and air that greatly increase border security and real-time targeting. Effective AI represents the competitive advantage across all organizational domains, military and civilian alike. When an AI-driven drone misses its target, the cost is incredibly high.  What is the risk to your organization when AI isn’t properly managed?   

The convenience of AI solutions may lead organizations to act without building the necessary guardrails, exposing themselves to significant risks. Knowing when to leap forward and when to pause for careful assessment is among the most critical ethical considerations many organizations currently face.

The rapid pace of AI development may easily push organizations to deploy solutions without rigorous oversight. AI can provide quick, cost-effective solutions, but that same speed can lead to errors, additional expense, and reputational damage. 

The accuracy and bias of data within AI models is well documented and cause for concern. If the underlying data is skewed, as seen in a 2023 incident where Gemini’s earlier model inaccurately depicted historical figures, organizations may face significant backlash from users. Generative AI’s ability to produce content like deepfakes only amplifies these risks. Like the internet, which has demonstrated its capacity for legitimate opportunity and exploitation, AI today offers both promise and potential peril.

Other ethical dilemmas emerge: Are organizations creating short-term solutions to long-term problems? Do AI models introduce vulnerabilities by relying on sensitive data, and how do leaders weigh the risks of convenience and cost? Managing these trade-offs is now a top priority for companies navigating AI adoption.

The Risks of Implementing AI Without Adequate Oversight

AI’s potential for rapid deployment can be both a blessing and a curse. Businesses that chase speed without establishing robust boundaries risk serious reputational damage, financial loss, or regulatory penalties. Two recent cases illustrate these pitfalls.

In 2023, Avianca Airlines found itself in a legal controversy when a lawyer representing the airline submitted a court brief generated by ChatGPT. The document referenced entirely fictitious case law, forcing the lawyer to admit the error in court and face professional embarrassment.

Similarly, CNET’s attempt to streamline content production with AI-generated articles backfired when the published pieces contained numerous factual errors and lacked proper citations. The backlash not only damaged CNET’s credibility but highlighted the risks of relying too heavily on AI without human oversight.

These examples underscore a critical truth: AI is not a silver bullet solution. Organizations need clear policies and expert oversight to avoid turning well-intentioned AI initiatives into public failures.

Balancing Innovation with Security and Privacy

The challenge of balancing AI innovation with data security varies across sectors but is essential to maintaining trust. Companies that fail to protect sensitive data risk catastrophic consequences. For instance, T-Mobile’s recent data breach settlement underscores the risks companies face without robust cyber measures. Similarly, National Public Data filed for Chapter 11 bankruptcy after suffering a massive breach of over 2 billion records. 

These cases highlight the real-world cost of poor data management—and the importance of building security into AI initiatives from the outset.

AI-driven solutions can enhance security practices when deployed correctly. For example, IBM reports that integrating AI into data management systems reduces the average cost of a data breach by $2.2 million. However, success requires more than technology; it demands careful leadership. While business objectives must remain front and center, organizations need to prioritize security to avoid costly mistakes down the road.

With regulatory scrutiny increasing across industries, companies must take a security-always approach. Financial institutions, food service businesses, and other sectors each face distinct regulatory frameworks, but all organizations share a responsibility to safeguard data. Ongoing engagement with regulatory bodies ensures that businesses stay ahead of new requirements while demonstrating accountability to customers and stakeholders.

Governance and Ethical Oversight: A Framework for Success

Effective AI adoption requires more than just technical tools—it demands clear governance frameworks and transparent oversight mechanisms. The following best practices can help leaders manage AI responsibly:

• Establish a cross-functional AI ethics committee. This committee should include representatives from various departments—such as legal, cybersecurity, marketing, and operations—to align AI initiatives with business objectives and ethical values. Including security professionals ensures the committee can identify potential vulnerabilities.
• Develop AI governance frameworks similar to those used for other transformative technologies. These frameworks set boundaries for AI use without stifling creativity. Organizations can consult resources like Palo Alto Networks’ AI governance guide for additional guidance on developing their policies.
• Promote transparency and invite scrutiny. Labeling AI-generated content, disclosing oversight procedures, and encouraging internal dialogue foster trust and continuous improvement. Organizations like iStock already use this approach, tagging AI-generated content to enhance transparency.
• Set explicit policies for sensitive data use. Policies should align with or expand existing security protocols, ensuring that AI applications don’t introduce new risks.

The key to effective AI oversight lies in metered and managed use. AI should complement existing resources rather than replace them entirely. Success depends on continuous human review by subject matter experts, which can prevent missteps and maintain organizational credibility.

The Role of Regulatory Collaboration in Shaping AI Ethics

Collaboration with regulatory bodies and industry groups is an essential element of responsible AI adoption. These partnerships provide businesses with early insights into new policies and help shape ethical standards. Organizations that actively engage with regulators not only stay ahead of compliance requirements but also influence the policies that define the future of AI.

Involvement with regulatory bodies offers several advantages. Threat data becomes more readily available, mitigation strategies are developed collaboratively, and security standards are improved for all. Businesses that fail to engage with regulators risk being caught off guard by new policies or public backlash, which can damage reputations and disrupt operations.

The adoption of generative AI offers immense potential, but it must be managed with foresight and care. Leaders who establish effective governance frameworks, engage with regulators, and foster transparency will be better positioned to unlock the benefits of AI while minimizing risks. The path forward requires thoughtful oversight, continuous learning, and a commitment to ethical responsibility—because in the age of AI, trust is as important as innovation. 

A seasoned leader and strategist with over two decades of experience in cyber operations, diplomacy, and project management, Dr. Briant Becote is currently a Cybersecurity Professor at the University of Advancing Technology, dedicated to student success in cybersecurity and computer science courses. He holds a Ph.D. in Cyber Operations and maintains PMP and CISSP certifications.