Finally it’s Friday: Is LinkedIn’s Auto-Tag Feature on Drugs?

Whatever AI LinkedIn uses for automatically associating “tags” to “posts” or “shares” is often impressively accurate (then again, I would guess that most of the “tags” really are just keywords extracted from the post/share, although these examples below most definitely are not). But, sometimes, it seems that that AI is just on drugs. Here’re a couple of examples of hillarious tagging of some of my posts:

 

LinkedIn Auto-Tag AI on Drugs: How do Datacenters deserve a tag of "Mechanical Engineering"?

LinkedIn Auto-Tag AI on Drugs: How do Datacenters deserve a tag of “Mechanical Engineering”?

This next example is equally hillarious. To a posting complaining about incessantly generalizing from (network) protocol comparisons, apparently LinkedIn’s AI thought “Ohhhh, Economics”:

LinkedIn Auto-Tag AI on Drugs: How does a post about protocol comparisons relate to "economics", "distributed computing" and "complexity"?

LinkedIn Auto-Tag AI on Drugs: How does a post about protocol comparisons relate to “economics”, “distributed computing” and “complexity”?

Well, maybe the AI just saw “curves going up and down” and concluded that Economics might be the right field? 🙂

Mostly, thought, the auto-tag feature is impressive. But, it’s hillarious when it’s not.