Just lately I, like so many people, am scratching my head to get a handle on AI, which I prefer to think of Assisted Intelligence. For me, the question I am asking is, how can we provide value to our consumers? At present we don’t seem to be doing anything that directly helps the customer, other than crippling and frustrating chat bots.
The rapid advancement of AI has introduced a paradox for knowledge workers. While AI systems can process vast amounts of data, there’s a disconnect between this capability and the practical tools available to professionals especially in Aviation and Travel handling both internal and public data.
Traditional large language models (LLMs) excel at high-level data assessment but often fall short in delivering actionable insights tailored to specific organizational needs. This rarely translates into consumer value. This gap highlights the necessity for practical AI appliances – both physical and virtual – that can seamlessly integrate and process internal and external data sources FOR the betterment of the consumer during the lifecycle of their interaction from ideation through to taking a trip.
My opinion is that current AI solutions are predominantly designed for broad applications, lacking the specificity required for travel, let alone individual organizational contexts.
This misalignment results in knowledge workers spending excessive time customizing generic AI outputs to fit their unique data environments, thereby diminishing productivity gains. The challenge lies in the absence of AI systems capable of understanding and processing proprietary internal data in conjunction with publicly available information.
For Aviation and Travel that is especially hard since much of our “knowledge” is both siloed and unfortunately our industry is not good at sharing.
An emergence of knowledge appliances?
To bridge this gap, the development of AI appliances – dedicated systems designed to handle specific data processing tasks – needs to be and appears to be gaining traction. These appliances can be tailored to align with an organization’s data infrastructure, ensuring that AI outputs are relevant and actionable.
Virtual AI appliances, deployed within existing IT environments, offer flexibility and scalability, enabling organizations to integrate AI capabilities without significant hardware investments. There are several – but not that many specific applications.
Most of these are being used to improve organizational efficiency in such things as Route planning for airlines and Pricing algorithms (which is a dark art at the best of times). There has been a concentration on chat bots which in the main I find annoying and unhelpful.
However, there are few examples in Travel Planning such as Romie from Expedia and from that data monster Google comes Gemini. At the lower level in specificity, I saw an example from start-up Wolfpack.
As for appliances – some have appeared and hopefully have died merciful deaths. The Rabbit R1 is an example. (And yes I did get one – very disappointing, I didn’t want to even try Humane Pin). I do find peripheral devices such as Meta’s Glasses to be a good example of how we can enhance consumers’ travel experience. Will our phones be the only place for such a physical or virtual knowledge appliance?
Future scenarios to keep an eye on
The convergence of open-source AI models and low-cost computing power both based in hardware or via the Cloud especially from the big AI players is likely to yield several developments. Examples are a possible proliferation of these devices.
I do not think that the there will be a rush to develop tools in the open to help companies to develop solutions – I really hope I am wrong about this. With the massive investment in large scale AI, there will be non-specific services that can be adapted.
I do believe there will be some AI enabled devices using ARM architecture as a low-cost variant of Nvidia. Even Amazon has announced it will get into that Chip business. As large corporations adopt AI as the norm, that will filter into general use. Also, there are some real AI tools arriving.
(However, there is a lot of nonsense of companies claiming to be “AI Powered”). One area to pay close attention to is Privacy and Security. There is an emergence (as the law struggles to keep up) of bidirectional data concerns. Not just the protection of data but its free use.
The need
The current disconnect between AI capabilities and practical applications for knowledge workers underscores the need for AI appliances that can effectively process both internal and public data. Collaborations between companies like Meta and ARM are paving the way for such solutions, leveraging open-source AI models and efficient hardware architectures. As these technologies mature, we can anticipate a more integrated and effective AI ecosystem that directly addresses the needs of knowledge workers. How can we in the travel sector turn this into value for our consumers and ultimately to “decomplexify” the products and services we offer.
Sources: I recommend being better armed with more information and reading the underlying sources. Here are some suggested reading for you:
Meta and ARM Collaboration on AI Models
Source: Thenextweb
Amazon Web Services (AWS) “Ultracluster” and AI Infrastructure
Source: The Wall Street Journal
Apple’s Privacy-Centric Approach to Generative AI
Source: Wired
ARM’s Vision for Scaling AI through Collaboration and Open-Source
Source: ARM Newsroom
Expedia’s Romie announcement
Source: Expedia
Meta AI Glasses
Source: Meta
Rabbit R1 hands on review
Source: Mashable
