The incentive for starting with AIR now is much less. Most if not all of us got started when Flash in the browser was still a viable platform.
But Flash player is being retired soon. It’s not been a viable platform for years, as casual gaming‘s moved to mobile while non-game web content (adverts, videos, websites) switched to HTML/Javascript.
From the decade or more it was the best technology for web games, many of us have experience, perhaps existing code, maybe whole web apps we want to port or adapt. The deprecation of the classic API in favour of Stage3D made it harder, but Starling and other APIs take care of a lot of the low level details.
Lastly Adobe’s support for AIR has been lacklustre in recent years, dropping support for Starling and Feathers and falling behind with upgrades to the platform. This will hopefully improve with Harman, but they have yet to announce much beyond a belated transition to Android 64-bit.
This all makes it hard to recommend AIR to anyone starting out. At least wait and see what Harman’s plans for it are, before you decide to commit to it. But also look at other technologies like Unity, or coding in the native language for a platform like C++.