Understanding GenAI

Generative AI involves sophisticated algorithms capable of creating content, making decisions, or identifying patterns from large datasets. Innovations from tech giants like OpenAI and Google have revolutionized tasks across various sectors, including hospitality. These technologies offer significant potential to personalize guest services and optimize operational processes both front and back of house. But they do require careful planning and implementation.

Practical Applications of AI in Hospitality

AI technology is poised to revolutionize both internal operations and guest-facing services in hospitality:

• Internal Operations (a few examples): In engineering, AI can predict maintenance needs, reducing downtime and operational costs. Marketing departments use AI to analyze guest data, tailoring marketing campaigns that significantly increase engagement and conversion rates. In human resources, AI assists with resume screening and predictive hiring, improving the quality and fit of new hires.
• Guest-Facing Services (a few examples): AI can enhance guest experiences through personalized room settings, automated check-in/out processes, and tailored recommendations for dining and activities based on guest preferences and past behavior. Dynamic pricing models powered by AI can adjust room rates in real-time based on demand, competition, and other market factors, maximizing revenue.
  ‍

Specific Risks Tied to AI Applications

While the potential uses are virtually infinite, at this stage, use of GenAI still presents several significant risks. These risks fall into the following major categories: privacy and data security, intellectual property, confidentiality, bias/discrimination, and accuracy.
‍

• Privacy and Data Security. In the hospitality industry, personalization and data analytics are becoming increasingly integrated into customer service and operations. This opportunity relies on collecting vast amounts of personal data from travelers. Businesses using AI to enhance guest experiences must ensure that they comply with data protection laws like GDPR by implementing robust consent mechanisms and transparent data policies. In addition, it is critical to ensure that appropriate measures are in place to protect the security of this data (e.g., encryption, access controls, anonymization where possible, regular security audits, etc.). These measures not only help in avoiding myriad legal repercussions but also build trust with customers, enhancing their overall experience and the value of the brand.
• Intellectual Property (IP) Issues: IP issues associated with GenAI are complex, particularly because they touch on both the input used to train these systems and the output they generate. There can be questions of ownership (because GenAI systems are trained on datasets that may contain the IP of third parties), as well as protectability (because content created by GenAI may not be protectable by IP laws). Understanding these concerns is crucial for organizations leveraging GenAI to ensure they comply with existing IP laws and mitigate potential legal risks.
• Confidentiality: Depending on what tools you are using, the data entered into the tool (the “prompts”) may not be confidential, and the same is true for the result (the “output”). Meaning, if sensitive company information is entered into a GenAI tool, that information may potentially be spit out in response to a query somewhere else.
• Bias/Discrimination: Bias and discrimination concerns with GenAI are significant (and the source of lawsuits already). They stem from inherent characteristics of the data used to train these models and the design of the algorithms themselves. These concerns can manifest in various detrimental ways, impacting individuals and groups unfairly and leading to broader social and ethical consequences.
• Accuracy: GenAI is only as good as the datasets on which it is trained. The models themselves may have issues that impact their performance (such as ‘hallucinations’). It is crucial for any user of GenAI tools to be aware that the outputs may not be accurate.

Engaging with AI Vendors
‍

When considering a new GenAI tool, or really, any tool that uses AI as part of its functionality, due diligence and clear contracting are key:

• Diligence: Ask a lot of questions about what is in the stack, how it is used, what the model was trained on, whether your data will be used to train the model, etc. Include infosec and privacy review as part of the evaluation of the AI tools.
• Define Scope and Performance Metrics: Clearly outline the services to be provided by the tool, expected performance standards, and metrics for measuring success.
• Address IP Rights: Specify the ownership of IP rights related to developed AI models and generated data.
• Include Compliance and Security Clauses: Ensure that contracts require adherence to industry-specific regulations and standards for data security.

AI Implementation Guidance Resources

There are an increasing number of tools available to help teams implement GenAI in smart, more secure manner:

• National Institute of Standards and Technology (NIST): Offers frameworks and guidelines for AI security and ethical considerations.
• Information Commissioner's Office (ICO): Provides advice on data protection and AI, helping businesses comply with privacy regulations.
• IEEE Standards Association: Develops globally recognized standards for AI systems, including ethical design and implementation practices.
• ISO/IEC JTC 1/SC 42: This is the joint technical committee of the International Organization for Standardization (ISO) and the International Electrotechnical Commission (IEC) that focuses on standardization in the field of artificial intelligence.
  ‍

Best Practices for AI Governance

To help manage these risks, organizations should be doing the groundwork now to adopt comprehensive governance practices, including:
‍

• Inventory current AI usage and evaluate areas of potential legal risk (such as from personal data usage, IP issues, bias, etc.).
• Develop an internal use AUP that addresses what tools can be used in what contexts, data that is appropriate for submitting, safety mechanisms to vet outputs, etc. Appoint stakeholder(s) that will consider new uses.
• Implement controls around internal AI use including technical measures to ensure compliance, privacy controls, model validation, human intervention/supervision, incident response and reporting and breach tabletop exercises.
• Update vendor approach (including due diligence and contracting processes).
• Update your customer approach (contracting processes).
• Invest in AI literacy and training; foster ethical AI practices.
• Conduct regular security assessments to protect against vulnerabilities in AI systems.
• Test AI systems regularly for accuracy and bias, especially those used in customer interactions and HR.
• Document all of the above in detail.
  ‍

As AI technology evolves, it offers both extraordinary opportunities and significant
challenges for the hospitality industry. By proactively managing the risks and
thoughtfully integrating AI into their operations, hospitality leaders can harness this
powerful technology to enhance efficiency and guest experiences while maintaining
compliance and ethical standards.

‍