NASHVILLE, Tennessee — Artificial intelligence has been a major buzzword in the hotel industry over the past couple of years, but there are still a lot of unknowns in regard to what it is, how soon it'll have an impact and how it can be applied across organizations.
During the "AI in Hospitality: How Will it Affect You?" panel at the recent Hotel Data Conference, Kurien Jacob, managing director and partner at Highgate Technology Ventures, said artificial intelligence and machine learning are really an extension of machines from the 1960s that were trained to do a certain thing.
"What I would say is you can't be fearful of it. It's actually going to help you improve productivity dramatically, and those who embrace it for that purpose versus saying 'Hey, is this going to replace what I'm doing?' Probably not. You're just going to increase productivity, and that's going to have a tremendous effect on the economy overall," he said.
Todd Brook, CEO and managing director at Unchained, said much of the apprehension that comes with technology comes from not understanding how it works.
One simple example of AI's applications is inserting a Word document or Excel sheet into ChatGPT and letting it read and interpret the files. Actually using the software is the easy part; compiling the data that goes into it is the hard part, Brook said.
Despite its similarities to machines of the past, there remains a certain amount of apprehension to fully buy in to AI. Arlene Ramirez, regional vice president of finance at CoralTree Hospitality, said the hotel industry shouldn't worry about it taking jobs away or changing how business is done, but should rather focus on how it will enhance productivity.
Lisa Targonski, director of commercial services at Elder Research, said AI allows people to upskill and get rid of mundane, repetitive work. In regard to fears of it taking over jobs, she pointed to other technological advances over the years.
"Calculators didn't remove the need to have physicists in our society, and the internet did not negate the need for teachers, and AI is not going to negate the need for human beings, especially in this industry," she said.
The fear of jobs being taken isn't completely unfounded, though. Jacobs said he believes once AI is commonplace, there won't be the same amount of people doing the same amount of work.
"We're going to have fewer people doing more things, because a big portion of what is being done will be done by AI, which gives you the time and ability to get focused on creative, strategic stuff, which is ultimately what's lacking," he said.
In the short term, jobs being replaced by AI isn't a concern, Brook said. Getting replaced by a human who has gotten themselves acclimated with the tools of AI is a more plausible concern.
"The big threat when I look forward to all of this, at least in the next two, three, four years, isn't necessarily the overall sense of job security," he said. "It's the idea of, are you going to be passed by with those that have taken that time to be just a little bit intentional of bringing this skill into their arsenal," he said.
Concerns and Risks With AI
One of the major concerns with generative AI is how one can trust the results that are spit out. Ramirez said it's a fair concern, but users have to first understand what they're asking from it in order to verify the results.
"When we ask questions of technology, we have to be very mindful exactly what it is we're asking," she said. "We have to understand our business in order to get those answers that will be valid."
Brook said in order to get the best responses, the minimum word count for a question should be about 200 words. No matter how precise the question, though, the No. 1 rule is individual accountability.
"It's done to make you faster, but you must check the work, you must be able to walk away — whether it writes something, it produces a result — you have got to check it, and you've got to do this many times until you are confident that this is a predictable output because it is really a representation of you," he said.
As companies begin to embrace AI more, several departments will use these tools. To mitigate risks such as uploading sensitive data or legal problems, companies should form an AI council, Targonski said.
"It's a really good time to start to think about pulling together an AI council to be thinking about these types of issues, because if marketing has one opinion and approach towards it, and sales has another opinion and approach towards it and finance has a third, you're going to end up with really disparate opinions and approaches across the organization and that's going to lead to problems," she said.
Implementing AI
Jacobs likened data to fuel to make an example of how artificial intelligence will work. To make fuel, one must first obtain oil and then take it to a refinery to clean and process it.
In the same way, users first need to collect the data and verify that it's correct before plugging it into an AI system.
"Otherwise, you're going to have an imperfect model of AI that you say doesn't work," he said. "That's what most people do. 'Oh, it just doesn't work.' But that's because the underlying data layer is not clean."
On the productivity front, Jacobs provided the example of putting together AI summaries every morning that can save 30 minutes of time each day. That time adds up and allows for more time to focus on the creative and strategic aspect of jobs.
Finding those applications and which AI tools fit in within a team is one of the first steps leaders must take, Ramirez said. There are a lot of different examples of AI, so some companies can only zero in on so much of it.
"How do you pack your suitcase if you don't know where you're going?" she said. "When we purchase technology or we're looking at our AI technology strategy, are you buying everything that's out there and trying to shove it in a suitcase that you're never going to wear, you're never going to use? We have to be very strategic about it, and just because it says AI doesn't mean that we need to buy it."