Generative AI and LLMs: What’s in it for the Tourism & Travel Industry?

It has been 4 months now since I've started walking on the path of understanding how to benefit from the latest developments of AI. I guess for those of us who are new to this, it's not an easy road as there's a lot of complex terminology and difficult to process fast developing innovations happening every single day. I've decided to take this path, because it's the future, even though we're on the tip of the iceberg of what's to come, missing out what's happening today could be disastrous in a not so far away future.

This article intends to go through what I've learnt until now and the challenges I believe exist for the tourism industry in the field of generative AI.

So, what's all the buzz regarding Generative AI?

What I've seen and read so far is that Generative AI is "a field of artificial intelligence that focuses on creating models and systems capable of generating new content, such as images, text, music, and even video". Unlike traditional AI models that are trained to classify or predict existing data, generative AI models aim to produce original and creative outputs based on the patterns and information they have learned from a given dataset. Generative AI is a rapidly developing field with the potential to revolutionize many industries, including tourism and travel.

Some of the developments in generative AI that are already impacting tourism and travel are:

 Virtual travel experiences, where users can explore destinations without ever leaving home, which is a great way for travelers to plan their trips, learn about different cultures, or simply experience the world from the comfort of their own homes.

Personalized recommendations to help travelers save time and money by ensuring that they are only considering options that are relevant to their interests and needs.

Automated customer service: to answer questions, resolve complaints, and provide support. This could free up human customer service representatives to focus on more complex tasks, such as providing personalized recommendations.

Optimized operations such as hotel bookings, flight scheduling, and transportation logistics. This could help businesses save money and improve efficiency.

Without doubt destinations will benefit a lot with these developments but

¿can destination management offices (DMOs) implement this type of technology?

Indeed they can. What we've seen as possible developments and solutions for businesses can also be scaled to a destination, for example:

Create virtual tours to promote the destination to potential visitors, help them plan their trips, or simply give them a taste of what the destination has to offer.

Generate personalized itineraries for potential visitors. These itineraries can be tailored to the visitor's interests, budget, and travel style.

Provide customer support: to support potential visitors. This can include answering questions, resolving complaints, and providing suggestions.

Optimize marketing campaigns: include targeting the right audience, creating the right message, and measuring the results of the campaign.

Improve data collection and analysis to better understand their visitors, their needs, and their preferences. 

By now most of you who have been following the developments of ChatGPT and Generative AI have witnessed its capabilities and how it can benefit tourism, but then another terminology appeared (at least for those of us that are new to all of this): LLM and with it how they play a significant role in the developments of what we've been seeing.

¿What is an LLM?

 LLM stands for Large Language Model. An LLM is a type of artificial intelligence that can generate text, translate languages, write different kinds of creative content, and answer questions in an informative way, even if they are open ended, challenging, or strange. LLMs are trained on massive datasets of text and code, and they can learn to perform a variety of tasks by analyzing the patterns in the data.

 LLMs are today still under development, but they have the potential to revolutionize tourism and travel. What we've seen until now is in the end thanks to the use of LLMs. The question is ¿where will this take us? I guess we’ll have to wait (but not for long) and one thing is for sure it’s going to be huge!

All aboard.. the LMM train!

Now, if a business want’s to jump on, where should it start? Well I guess it’s important to start with understanding and learning about available LLM Models. There are a number of them available today, the most popular ones are:

• Bard (Google AI)
• GPT-3: (OpenAI)
• LaMDA: (Google AI)

The listed LLMs have been trained on a massive dataset of text and code. They can all generate text, translate languages, write different kinds of creative content, and answer your questions in an informative way.

Could destinations develop their own LLM?

I've been reading a lot about LLMs and how today anyone can have their own local AI or private chatGPT even without having to be connected to the internet, which I find mind blowing! One of the "gurus" I follow in this development is Brian Roemmele of Multiplex. He has been the past few months working with other likeminded masterminds in developing personal AI on local PCs where they can interact with local data that exist in a particular PC, in other words personal data you don't want anyone to see or corporate you prefer not to share with the cloud. This type of development is just weeks old, improving by the hour with thousands working on how to improve these open-source and free AI platforms.

With the idea of having personal chatGPTs on local PCs, I started to consider the possibilities this could mean for a country DMO and the benefits it could mean if they were to develop or train their own LLM or their own chatGPT.

¿But why would a business or DMO want to have its proper LLM?

Here are some reasons that can justify such an idea:

Increased personalization: An LLM that is tailored to a specific business or organization can be used to create more personalized experiences for visitors/clients. This could include things like generating custom marketing materials, providing personalized customer service, or creating unique learning experiences.

Improved efficiency: An LLM can be used to automate tasks that are currently done manually, such as data entry, customer service, or marketing. This can free up employees to focus on more strategic or creative work, and it can also help to improve the accuracy and consistency of these tasks.

New opportunities: An LLM can be used to explore new ideas and possibilities that may not have been considered before. This could lead to the development of new products or services, the improvement of existing ones, or the discovery of new ways of doing things.

Now, to get to the point of having a local and self made AI, a DMO would need to:

1. Collect a massive dataset of text and code. This dataset would need to be large enough to train the LLM on a wide range of topics associated with the tourism activity of that destination and information that the destination could consider relevant for the type of analysis it wants to undertake.
2. Train the LLM on this dataset. This process would require a lot of computing power and time.
3. Fine-tune the LLM for its specific needs. This would involve providing the LLM with additional data and feedback.

Once a destination (or business) manages to develop its own LLM, it would be able of valuable use as it could perform a variety of tasks, such as:

• Improve its marketing campaigns.
• Develop new products and services.
• Increase tourism revenue.
• Generate text, such as blog posts, articles, and social media posts.
• Translate languages.
• Write different kinds of creative content
• Answer questions in an informative way.
• Provide personalized recommendations.
• Automate customer service tasks.

However, it is important to note that developing an LLM is a very complex and expensive undertaking. It is not something that most destinations would be able to do on their own. Instead, they would need to partner with a tech company that specializes in developing LLMs. This would mean a significant number of issues and costs that a destination would have to consider and that most would probably opt out before even considering it. For example:

The size and budget of the DMO: Larger DMOs with larger budgets may be more likely to be able to afford the costs of using generative AI.

The expertise of the DMO's staff: DMOs with staff who have expertise in data science and machine learning may be more likely to be able to successfully use generative AI.

The willingness of the DMO to experiment: DMOs that are willing to experiment with new technologies may be more likely to benefit from using generative AI.

High cost: Developing an LLM is a very expensive undertaking. It requires a large dataset of text and code, as well as a lot of computing power and time.

Data collection and preparation: DMOs will need to collect a large dataset of text and code to train a generative AI model. This data can be expensive to collect and prepare.

Model training: Generative AI models require a lot of computing power to train. This can be a significant cost for DMOs that do not have their own data centers.

Model deployment: Once a generative AI model is trained, it needs to be deployed so that it can be used by DMOs. This can involve the development of a user interface and the integration of the model with the DMO's existing systems.

Maintenance and updates: Generative AI models need to be maintained and updated on a regular basis. This can be a recurring cost for DMOs.

 The cost-benefit of using generative AI for destinations depends on a number of factors, including the specific needs a destination may have and the availability of resources. However, the potential benefits of using generative AI can be significant, and destinations should at some point consider the use of this technology.

LLMs and the potential risks for the Tourism Industry: Data privacy concerns with implementing and use

Perhaps one of the biggest limitations/issues today for the deployment and scale up of this technology in tourism (and other industries) is linked to the potential data privacy concerns associated with the use of LLMs. Every level and category of how travel businesses and destinations work with data has its issue:

• Data collection: LLMs are trained on massive datasets of text and code. This data can include personal information, such as names, addresses, and phone numbers. Here Businesses need to be careful about how they collect this data and that they have the consent of the data subject to collect and use this data.
• Data storage: LLMs require a lot of data to train and maintain. This data can be stored on servers that are owned by the tourism business or by a third-party vendor. In this category, it's necessary to take steps to protect this data from unauthorized access, use, or disclosure.
• Data use: LLMs can be used to generate text that is similar to human-written text. This text can be used for a variety of purposes, including marketing, customer service, and research. The use of data has to be done securely and businesses/destinations have to reassure that they will not use it in a way that violates the privacy of the data subject.
• Data sharing: LLMs can be shared with third-party vendors. This sharing can be done for a variety of reasons, such as to improve the performance of the LLM or to develop new products and services. Destinations and businesses have to make sure that they do not share it with a third-party vendor that does not comply with the same privacy standards.

In the case that destinations and businesses consider using LLMs, they will need to take steps to address these data privacy concerns:

• Obtain consent
• Encrypt data
• Limit access
• Implement security measures:
• Train personnel
• Provide transparency
• Provide opt-out

Conclusions

LLMs are a powerful tool that can be used by tourism businesses to improve their marketing campaigns, develop new products and services, and increase tourism revenue. However, there are some challenges that need to be addressed before LLMs can be widely adopted by tourism businesses.

One of the biggest challenges is data privacy. LLMs are trained on massive datasets of text and code, which can include personal information. Tourism businesses need to be careful about how they collect, use, and share this data. They also need to be transparent about their data privacy practices and give visitors the opportunity to opt-out of data collection.

Another challenge is cost. LLMs can be expensive to develop and use. Tourism businesses will need to carefully consider the cost-benefit of using LLMs before making a decision.

Finally, tourism businesses need to have the technical expertise to use LLMs. LLMs are complex technology, and tourism businesses may need to partner with a technology company that specializes in developing and using LLMs.

Despite these challenges, LLMs have the potential to revolutionize the tourism industry. Tourism businesses that are willing to address these challenges can reap the benefits of this powerful technology.

Overall, LLMs have the potential to significantly impact the tourism industry. Destinations and businesses that embrace this technology will definitely develop competitive advantages and lead the way for the coming years, it will depend on those (us) that work in this industry how far and deep we want this technology to permeate, penetrate and transform the way we work and how we connect and inspire those that just want to paradoxically disconnect from everything and enjoy travelling the world.

#generativeAI #tourismAI #HospitalityAI #chatGPT #LLM #BARD #GPT3 #tourismconsultant #AI #IA