More on the Legal Implications of AI for Travel Advisors

In keeping with the extremely rapid spread of AI technology, I have had some new insights since my October 16 article on this subject. My list of possible uses/misuses should include additional items. To approach this vast subject in a manageable way, I have added to and reordered the list:

Setting Up the Business

• Determining the legal form of business for an advisor/agency enterprise
• Deciding what insurance and licenses are required for being an advisor or agency
• Creating a business plan
• Evaluating legal issues generally

Managing the Business

• Hiring/firing decisions
• Assessing the implications of using employees vs independent contractors
• Correspondence with clients
• Analysis/comparison of trip options
• Correspondence with suppliers

Promoting the Business

• Marketing materials
• Press releases
• Destination information
• Original images without a camera
• Website content
• List making
• Creating search terms for the Internet

Evaluating Business Performance

• Comparing to the business plan
• Comparing against specific competitors
• Relying on AI-generated reviews
• Comparing against industry data

Acquiring/Disposing of the Business

• Evaluating whether expansion through acquisition is warranted
• Determining the value of the business for purposes of merger/sale

To begin at the beginning, when considering the opening and legal framework for a travel advisory business, it may be tempting to use AI-based tools to avoid paying legal fees for at least the following essential actions:

• Determining the legal form of business for an advisor/agency enterprise
• Deciding what insurance and licenses are required for being an advisor or agency
• Creating a business plan
• Evaluating legal issues generally

Don’t do this. There are AI-based tools that provide some legal information and that can perform some legal analyses, but (1) they require legal training for their proper use, and (2) they are limited as to the tasks they can perform properly. The creation of a new business with proper legal formalities to protect the owner from liability and comply with state and federal laws is a step of huge importance and should only be undertaken with advice and, depending on the complexity of the business setup, active engagement of qualified counsel.

My view of this applies to all four of the categories identified above. While the competency of AI-based tools is growing rapidly, these fundamental startup issues are beyond the scope of such tools. A quick read of the terms/conditions and warnings in those technologies should confirm what I am saying.

Starting a business with a solid legal foundation and a well-considered business plan specific to the circumstances the business faces is simply too important to justify reliance on machine learning no matter how seemingly advanced.

The second category of subjects list above is Managing the Business,” where I have identified five categories in which AI-tools may be thought useful.

• Hiring/firing decisions
• Assessing the implications of using employees versus independent contractors
• Correspondence with clients
• Analysis/comparison of trip options
• Correspondence with suppliers

My views here are the same for the most part.

Hiring/firing decisions are notoriously difficult, involving subtle and highly judgmental decisions about the past and future performance of personnel. Because they involve another party who is often in a highly dependent relationship with the business, these are among the most sensitive and fraught with peril decisions that any business owner will make. I am aware of the reports that Human resources managers use machine learning software to analyze millions of employment-related data points, churning out recommendations of whom to interview, hire, promote, or help retain.

This is happening but it would be a mistake to believe that the decisions being made are the “best” decisions for the firms involved or that they are doing justice to the qualifications of the applicants being judged, and disqualified, in this manner.

These concerns are especially serious for another reason: the uncertainty and change around the issue of which workers are employees and which are independent contractors. The law governing those issues was changed by the Trump 2020 administration, changed back by the Biden administration and we are now on the cusp of a resumed Trump presidency that may lead to further changes. These are challenging issues in the best of circumstances, but in this chaotic situation it is extremely unlikely that AI-based technologies can provide reliable direction.

On the other hand, correspondence with clients may represent a partial, if limited, exception. AI tools have sufficient capabilities that, if provided with proper prompts, may be helpful in saving time by creating drafts that, following review, will net some time-saving. I emphasize “review” because, as the AI tools themselves repeatedly warn, they often make mistakes. Moreover, no matter how “real” they may seem, AI tools are not human, do not have human experiences and other human qualities, and do not know much if anything about the agency’s clients. It is therefore imperative to avoid relying on AI for client correspondence unless and until it has been reviewed by someone in the firm who does know the client and the reasons for the correspondence.

The same issues arise regarding the analysis of trip options. Being unaware of the client’s expressed concerns, including past travel experiences, an AI tool is at best a crude device for making client travel recommendations, regardless of how precise the prompts are. Indeed, the requirement of developing client-specific prompts may introduce more time and complexity into the process when the advisor who knows the client is already armed with the necessary information and judgment on which to base recommendations. All AI-generated information must be reviewed before placing it before a client.

Finally, every advisor understands the importance of maintaining relations with suppliers. Sending them inappropriate communications can damage credibility in ways that are difficult to recover. So, again, review, review, review anything created by AI that is intended for third parties.

In the next article in this series, we will examine the remaining issues in our list of concerns.