Humans and animals with eyes are pretty good at quickly identifying things seen in the past. It is a basic survival skill. “Identify the person”, “Identify the weapon”, “Identify the ripe fruit”.
So when training populations for basic survival, you probably have to go back to “Identify and classify quickly” – if you value humans doing a particular job.
Go through all the movies you have seen in your life, all the games you have watched, all the classrooms you have seen – and look at those places where the humans involved were praised for getting answers quickly, winning the prize. Because they were fast classifiers. Look at symbolic mathematics and computer programming – moving symbols around.
Emoji’s would be OK, if they also can quickly be explained – when a new one is presented. The computer could keep track of what each person has seen, and when a new emoji or image comes to a person’s screen or vision — identify it and explain it.
Those lookups where you type and it puts the related materials and previous searches, and most likely uses – those work fine for characters, emojis, images, icons, diagrams, patterns, colors, textures – things that people can see and identify, things that people can remember, recombine and share.
I have studied all the symbols used on the Internet. An emoji language is certainly possible. But people have different skills in seeing, so enlarging and changing the colors and contrast – are needed too. Then you just get to an image tool that lets you place images and copies of images onto a screen or into a 3D framework – for humans to work with those images, change and rearrange them – and share them. That is pretty close to art, design, visualization, science, technology, engineering, communication.