Keeping the "Human in the Loop" in travel

Since the release of ChatGPT and other natural language processing (NLP) and large language model (LLM) tools, there has been a never-ending stream of articles and information about artificial intelligence in our industry. This advancement has been in the works for years. However, it is new to many of us, and the development of this technology has set the world on fire.

Albert Einstein and Isaac Newton have already said, “What we don’t know is much more than what we know,” and “What we know is a drop. What we don’t know is an ocean.” So with those words of wisdom, it is critical for us to take a beat, learn all we can, and try to understand the implications of this new technology — at least as best we can today.

HITEC 2024 was abuzz with the potential uses of NLP; all the while, Google, Microsoft, Adobe and many more technology giants have already implemented and integrated NLP into their software. If you use any of these platforms, you may have seen the transformation as search engines reveal more relevant data, cloud-based productivity platforms seem to know what you want to say before you say it (and often it certainly says it better!) and graphic platforms are leveraging works in progress to teach NLP platforms. 

With all that going on in the background, it can be a slippery slope for data providers to ensure the protection of personal data or intellectual property. For creatives, we are already seeing a shift. We have seen companies like Adobe update the terms of service for their generative AI products, stating that Adobe may use techniques such as machine learning to analyze user content to improve its services and software.

While the reaction to this has been mixed at best, the reality is that we will never know the exact nature of how NLP is being integrated into our daily lives unless we are part of the changes being implemented. What we can see is the results and, hopefully, how those results make our lives better. 

What we do need is assurance that the advancements in development work for the human good. To do that, we require humans to play a key role in these developments. 

What is human-in-the-loop in AI & ML?

Many of us in the hospitality industry have never heard of this term. I read an interesting article written long before ChatGPT made its appearance in the market. Here’s a brief primer.

Human-in-the-loop (HITL) machine learning is a collaborative approach that integrates human input and expertise into the life cycle of machine learning (ML) and artificial intelligence systems. Humans actively participate in the training, evaluation or operation of ML models, providing valuable guidance, feedback and annotations. Through this collaboration, HITL aims to enhance the accuracy, reliability and adaptability of ML systems, harnessing the unique capabilities of both humans and machines.

While ML models possess remarkable capabilities, they can benefit from human expertise in areas requiring judgment, contextual understanding, and handling incomplete information. HITL bridges this gap by incorporating human input and feedback into the ML pipeline. 

This human collaboration enhances adaptability and allows models to evolve with changing user preferences and real-world scenarios. By integrating the human element, we empower ML systems to navigate the complexities and nuances that often challenge purely algorithmic approaches. This pairs our ability to contextualize, think critically and sift through the noise with algorithmic machine learning models’ incredible ability to process and quickly synthesize huge amounts of data so the strengths of both humans and machines shine through.

The importance of HITL in revenue management

There is a lot of concern that AI and NLP advancements will replace jobs. Just as the internet, the cloud and mobile devices have changed our lives and replaced the way we used to do things, AI and NLP will advance our society further, and it seems much faster. For revenue management professionals, it will be incredibly important to stay connected to the data and rationalize/explain the outputs. 

While they are great “copilots,” AI-based systems do not have all the answers. As a human revenue manager, it is critical to be able to question and/or challenge the data or outputs and validate accuracy and relevance. In fact, revenue managers have a great advantage with AI since they are already data custodians and have been using AI-driven RM systems.

Will people lose their jobs because of AI? No, but they will lose their jobs to people who know how to use AI tools and systems most effectively. That is no different than someone not knowing how to use a property management system to check-in guests or refusing to engage with email – it is the future, and the more people embrace it, the more they will be equipped to take advantage of it.

We will see AI architects in the future who will look at the right utilization of AI across an organization — how can AI be deployed to be most effective across the entire business? What insights can be derived from a system that eventually has access to all accessible data? These are questions that remain to be seen, but certainly there are many clear benefits to keeping humans in the loop. 

Enhanced accuracy and reliability require human input and oversight to significantly improve the accuracy and reliability of ML models. Bias mitigation needs human involvement to help identify and mitigate potential biases in data and algorithms, promoting fairness and equity in ML systems. 

Increased transparency and “explainability” are crucial. Human insights help explain behind-model decisions, enhancing their transparency and interpretability. This also improves user trust. The inclusion of human feedback and collaboration fosters trust among end-users, increasing their confidence in ML systems.

Finally, continuous adaptation and improvement are necessary. Feedback gathered during HITL serves as a valuable source for ongoing model improvement and adaptation to evolving real-world conditions. 

Will there be a time when humans aren’t required?

Remember the advent of the internet, email and cloud computing? Remember the pains we went through to understand these advancements? Did we understand how these technologies would change our lives? Certainly not. Who knew then that we could order food, a ride or check our home security alarm from another location through a device in our pocket? Who knew we could go online and order anything from anywhere at any time? 

The entire premise of AI and NLP is to help humans be more productive and efficient. But with great change comes great responsibility. Data companies are leveraging these advances to ensure users can interact with data more easily and quickly. It removes a lot of the “button pushing” and changes our relationship with data. It will be imperative to build safety protocols to protect sensitive and proprietary data.

The list goes on, but when we look back, technological advancements have been moving forward full steam ahead for decades. The emergence of a tool that speaks our language shouldn’t surprise us. Those who adopt it, learn to use it and engage fully with its potential will be the game changers and innovators of tomorrow.