Why Hospitality is the Ultimate Career Path – Best AI-Generated Hotel Industry Recruitment Ad Yet

Last summer, I created a couple of experimental hotel industry recruitment ads (see links below), mainly to test Midjourney image generation capabilities, plus ElevenLab’s multilingual audio features.

I Created The Very First AI-Generated Hotel Industry Recruitment Advert

Hotel Industry Recruitment Ad: A Second Take Using Multilingual AI Audio

We know Generative AI text to image generation quality has come a long way…in fact it is pretty difficult to tell real from AI generated images much of the time now. Videos, not so much. But things HAVE improved a fair bit in the last 12 months. A great example is the latest output we’re seeing from the likes of RunwayML, Kling, Luma Labs and others (including those not widely available, like Sora). Here’s a travel ad I created recently, using Runway’s text to video capabilities.

Creating The First AI Travel Ad For Less Than $100 – Visit India, The Real Life Jungle Book

With RunwayML finally releasing its image to video feature on Gen3 Alpha a couple of days ago, I decided to give the tool a test run, creating a brand new hotel industry recruitment ad…and I can confidently say that version 3 (vs versions 1 and 2 created last summer) is definitely the best so far*!

Here’s the final product:

*At least until Gen AI video capabilities improve even further.

How I Did It

Here are the tools and techniques used:

• Scriptwriting – The ideation and scriptwriting (including the narration) was done using ChatGPT.
• Visuals – The base visuals were all created on MidJourney (latest version launched recently is 6.1) and converted to video using RunwayML’s Gen3 Alpha’s image to video feature.
• Audio – I used ElevenLabs for the audio narration.
• Music – This was created using Suno AI.
• Editing – As usual, I used Final Cut Pro X…though pretty much any video editing tool would work equally well.

Notes and Observations

• ChatGPT remains a handy tool when it comes to ideation and script generation, as long as you take the time to give adequate context and iterative guidance. You still get the best output when you combine the LLM’s capabilities with your own intuition and creativity.
• Midjourney is still the Rolls Royce of text to image generation and it keeps getting better. There’s still a bit of wonkiness in some images, especially those featuring lots of people, hands, etc. but for the most part it does a pretty incredible job.
• RunwayML’s image to video feature with Gen3 Alpha is a game changer in terms of the visual control you can now introduce into your creations…a good example is the first few shots of the ‘younger me’ I created and animated using this tool (no longer are you confined to just Ken Burns and other video movement effects in your editor of choice!). Some of the video outputs still suffer from glitches (e.g. morphing, backward movement, distorted anatomy, etc.) and output is limited to 720p (HD) vs 1080p (Full HD), but many of the videos (or at least selected portions) can be used straight out of the gate.
• Suno is terrific for music – in this instance I simply provided the style of instrumental music I wanted and the tool did the rest.
• ElevenLabs is still the best text to audio tool out there, with an ever growing library of voices to choose from. I have my own voice on the tool too, but decided to go with a deeper voice from the library instead.

AI video is getting pretty good, though it still has some way to go…not bad for just a year, though. Can’t wait to see where we are in another year from now! What did you think of this latest experiment? Leave a comment or DM to let me know.