Artificial intelligence and machine learning open up many possibilities for improving business processes and taking advantage of the world of big-data. Your company has data from countless sources, whether you’re looking at transactional data or social media data, but it’s impossible for humans alone to look through all available information and create actionable insights from them.

That’s where artificial intelligence, specifically machine learning comes into play. Machine-learning algorithms can process your data at a rate impossible for a human. Once they have “looked over” the data, the software directs human attention to the right pieces of information to which they can apply uniquely human thought patterns, creativity and pattern recognition and pick out false positives.

The machine-learning solutions become more accurate over time, as they take what they learn from human or crowdsourcing interaction and apply it to the next process. The combination of humans and machines shows what a valuable collaboration this can be. The world doesn’t have true artificial intelligence yet, but machine learning is set on the right path.


Credit: CC BY-SA 2.0, 121483302@N02, Flickr

Who Uses Artificial Intelligence With Human Collaboration?

The technology not only sounds fascinating for artificial intelligence development, but it also has many practical applications in the current business world. You probably benefit from some of these human and technology combinations already and you don’t even realize it. Here are some ways companies are using machine learning and humans to improve their products and services.


Airbnb connects travelers and hosts to each other as a sharing-economy company. One of the concerns Airbnb deals with on a regular basis is fraudulent behavior, such as hosts accepting payment and canceling reservations or travelers creating multiple accounts to hide negative reviews. That’s why Airbnb uses machine learning to predict fraud activity, and has a trust and safety team looking through results to handle false positives and other data points.


Portuguese based Feedzai is a fraud protection platform built on machine learning. It uses artificial intelligence in its fraud-detection algorithms to help the system detect changing fraud patterns instead of operating on a system that looks at a set series of rules.

Rules-based fraud detection makes it difficult to keep up with sophisticated fraud methods used, while machine-learning systems can look at the data in context, and create new insights based on what it knows and what it learned from the humans using the system.

This platform is designed to streamline machine learning, so fraud and risk teams don’t need a data scientist to understand potential fraud alerts. They can look at the system and apply their unique understanding of fraud methods and potential risk to make decisions from there. The system looks at the team’s actions to improve its fraud detection performance.


PayPal is an online payment processing company used by millions of customers worldwide. This business is another one implementing machine learning into fraud detection practices and other forms of risk management. PayPal uses deep learning algorithms for machine learning and multiple methods for analyzing data to determine if a transaction is fraudulent, running linear, or neural net.

The result goes to PayPal’s data scientists for further evaluation. They take a critical eye to the results and enter additional information into the system. The scientists input the final evaluation results into the system, and also choose the data sets for all testing, so humans are a vital part of the process.


Unbabel delivers online translation services combining artificial intelligence and machine translation with a crowd on online translators; the Unbabel system learns from each addition to it’s machine translation done by human translators. Translation is more than simply creating a word-for-word translation of a written piece. Context has a large impact on the target language and human translators are better at understanding nuanced context, but the machine learning capabilities of the system that Unbabel uses helps them deliver consistency.

Over time, the intelligent translation service understands the right phrase to use in the right context, which speeds up how quickly human translators can complete their work. The system can point out potential translation problems and provide insight on how to fix issues to assist human translators (disclosure: Unbabel is a client of Blonde 2.0, the company I work for).


Captricity is bringing machine learning and crowdsourcing to data capture. The service is designed to quickly extract data digital documents to capture customer information without forcing an employee to manually enter all the data from the paper. Captricity uses a machine learning system to recognize characters, handwriting, and other information included on the forms once it digitally shreds the paper down. These results are verified and corrected by a crowdsourcing pool viewing the individual shreds of paper. Any results that humans rate “gold star” are used by the system to further its data recognition efforts.

The list goes on, many more companies are creating systems including machine learning backed by humans and crowdsourcing pools to help them speed up business processes, improve results, and create the closest thing to artificial intelligence available currently. These solutions show the true benefit of human collaboration with technology, adding unique insights and oversight to the processes. Humans have the tools they need to do their jobs faster, with less manual input, while the algorithm gets accurate and corrected data to improve the overall system accuracy.