Artificial Intelligence in Business: Five Legal Risks You Should Not Overlook

By Mario Alberto Arias V. · Founding Partner

Over the past few months, I have had more than a few conversations with clients, colleagues, and friends who enthusiastically share everything they are doing with artificial intelligence: contracts, proposals, presentations, marketing campaigns, financial analyses. And I welcome that speed. But increasingly, those conversations end with a word of caution — “Be careful, double-check that” — or an uncomfortable question: “Does this have legal implications?”

The short answer is yes.

AI has entered the business world faster than legal awareness of how to use it responsibly has developed. This has left those of us in the legal profession thinking, learning, and closely following developments in foreign jurisdictions that are already a step ahead.

Here are five issues that, in my view, every business should be considering.

1. Ownership of What AI “Creates”

In Panama, as in many other jurisdictions, content generated entirely by AI may not qualify for copyright protection. Your logo, advertising copy, software code, or illustrations may not legally belong to you in the way you assume.

Worse still, AI-generated content may incorporate third-party copyrighted material used during the model's training, potentially exposing your business to intellectual property claims you never anticipated.

2. Confidentiality and Data Protection

This is perhaps the most underestimated risk.

Uploading contracts, financial statements, client information, or personal data to a publicly available AI platform may breach professional confidentiality obligations, Panama's Law 81 of 2019 on Personal Data Protection, or confidentiality agreements your company has already signed.

In regulated industries — including banking, healthcare, financial services, accounting, and legal services — the risk of regulatory penalties is very real.

3. AI “Hallucinations” and Their Contractual Consequences

AI makes things up. This has been demonstrated time and again.

It fabricates case law, contractual provisions, figures, and citations.

When those fabrications find their way into a signed contract, a submitted quotation, a court filing, or a binding commercial communication, the company remains accountable for what it agreed to or represented — not for what the AI “intended” to say.

We have all encountered AI-generated errors that could have proven extremely costly had they gone unnoticed before execution.

4. Duty of Care and Fiduciary Responsibility

A director, executive, or licensed professional cannot delegate their judgment to a machine.

Investment decisions, hiring decisions, professional recommendations, or work product based exclusively on AI-generated output, without qualified human review, may constitute a breach of the duty of care and expose the decision-maker to personal financial liability.

Technology may support professional judgment, but it cannot replace the responsibility that comes with exercising it.

5. AI in Employment Decisions

Using algorithms to screen résumés, evaluate employee performance, or justify dismissals may sound efficient — until the algorithm produces a discriminatory or unreasonable outcome that your company must defend before the Ministry of Labor or a court.

Responsibility for employment decisions remains with the employer and the individuals making those decisions. Using AI does not transfer that responsibility to the technology.

It Is Not About Slowing Down. It Is About Governance.

Artificial intelligence is one of the most powerful tools we have seen in decades, and I am certainly not suggesting that businesses avoid it. Quite the opposite. We all use it actively.

What I am suggesting is that we stop treating AI as a corporate toy or a cure-all solution and start governing it for what it truly is: a technology that directly affects contracts, data, decisions, and people.

Three straightforward steps can make a meaningful difference:

1. Establish a clear internal AI policy defining what information employees may and may not upload to AI platforms.

2. Require qualified human review before signing, submitting, or publishing AI-generated content.

3. Carefully select enterprise-grade AI tools that provide appropriate confidentiality and data protection safeguards.

Speed is a competitive advantage. Speed without sound judgment is a problem waiting to happen.

At Arias, Ábrego & López, we assist businesses in developing artificial intelligence governance policies, protecting their data, and managing the legal risks associated with this rapidly evolving technology.

If you would like to discuss how to implement responsible AI governance within your organization, we would be pleased to hear from you at [email protected].