Toloka Global Community: Join, Learn, and Contribute

crowdsourcing
|
Mar 22nd 2022
Toloka Team
Image

Sure, there’s a lot to like about earning with Toloka. It’s safe, it’s clean, there’s no office politics. It pays on time and the hours are as flexible as it gets. And yet, you may wonder: why would anyone pay you to do these tasks?

It’s About AI

Artificial intelligence (AI) can do amazing things. Home assistants understand what you say, cars drive themselves, security systems recognise faces. You’d almost think it’s magic. But deep down, it’s all still a lot of tiny transistors passing small parcels of electricity to each other. Like every modern computer.

Where it differs is that traditional computer programs just follow instructions. The developers have to anticipate every circumstance, and tell the machine precisely how to respond. An AI though, is programmed to watch and learn. After a while, it can decide things on its own.

But It Needs Your Help

Right now, you’re reading these words on the screen. You recognize the letters. You’re probably so used to it, you don’t even think about it. But were you born like that? Nobody was!

As children, we were told about each individual letter, at school, on Sesame Street and through bedtime stories. Not once. Many times. Over and over again. After a while, we learned to connect each letter with a particular sound, and then to read whole words and sentences. An AI also needs examples to study. We call this training data. And just like teaching a child to read, it requires repetition and reinforcement. It takes a lot of data, and it needs to be correct.

Getting Training Data

There are so many places to find training data. You can scour it from the public internet. You can buy public datasets - or even get them for free. You can show the AI movies or have it listen to the radio.

For some tasks, though, none of these are quite good enough. When you want an AI to think like a human, the best data comes from watching real humans make decisions.

Labeled and Unlabeled Data

It also matters whether the training data is labeled. That means each item has a little note that tells the AI what to think. That label might identify the item: whether, for example, an image contains an apple or a cat. Or the label might say whether a decision was right or wrong.

An AI can be trained on unlabeled data, but this has its challenges and can mean more errors. Again, think back to when you first learned to read. Children’s books have lots of pictures to help with every word. Without those pictures, you still would have figured it out eventually. But it would have been much harder, with more mistakes along the way.

Building Things That Matter

Toloka tasks really aren’t all that different to teaching a young child an alphabet. It’s a hugely versatile source of training data. It can be used to generate totally new data. Toloka can also rapidly turn unlabeled data into labeled data. Toloka can even correct an AI’s mistakes so that it learns and grows.

That’s Where Toloka Comes In

All around the clock, our crowd of 200,000 monthly active Tolokers are building new technologies. Technologies that improve healthcare. That protect the environment. That make life easier and more convenient. Tolokers build the future. It’s an exciting thing to be part of.

So Try it, and See if You Like it

Toloka tasks really aren’t all that different to teaching a young child an alphabet. It’s a hugely versatile source of training data. It can be used to generate totally new data. Toloka can also rapidly turn unlabeled data into labeled data. Toloka can even correct an AI’s mistakes so that it learns and grows.

Building Things That Matter

If you want to do work that has a lasting impact at a global level… and make money along the way, give Toloka a try. You don’t need a degree in machine learning to take part. You don’t need to be any particular age, or live in the hip part of town. All you need is an internet connection and an attention span.

It’s totally flexible: you can work as much or as little as you like, at your convenience. And if you’re not sure it’s for you, just give it a try. If you change your mind, you can just walk away.

Similar articles