Let’s chat

Rewind: Top 4 AI Trends Predictions of 2020

Vadim Peskov
Vadim Peskov
Rewind: Top 4 AI Trends Predictions of 2020

Throughout the year, data and software security have become increasingly precious commodities.  As a result, cybersecurity underwent a major shift towards AI-based systems for recognizing and preventing common types of cyberattacks. In 2020, artificial intelligence also sparked an increase in automation in various industries.  

Almost a year ago, our blog post predicted the 4 emerging artificial intelligence trends of 2020. It’s now time to rewind and see if our predictions of AI trends came true.

  • Improvement of dialog capability and smarter context handling in AI assistants

Conversational artificial intelligence refers to technologies, like chatbots or voice assistants, which users can talk to. According to Statista, 4.2 billion digital voice assistants were being used in 2020 in devices around the world. 

With the growing expectations from voice assistants, AI and machine learning algorithms were  developed to understand users’ search behavior and context significantly better. 

Overall, companies will continue working on enhancing the voice assistant’s capabilities of voice recognition and overall contextual comprehension to make the experience more accurate to human interaction and helpful for the user. 

  • AI will work with smaller datasets

For some companies, big data is simply not available or too expensive to obtain. In 2020, some of the solutions for leveraging deep learning, even with limited data, were proven to be effective, including, ‘fine-tuning’, the idea of taking a very large data set that is hopefully somewhat similar to your domain, training a neural network, and then fine-tuning this pre-trained network with your smaller dataset. To get started with some code for fine-tuning, check out Pytorch’s tutorial.

For those who don’t have any success with fine-tuning on large datasets, data augmentation is the next best bet. The idea behind data augmentation is simple: change the inputs in such a way that provides new data while not changing the label value. We saw a few successful examples of these technologies in our projects. 

  • Eliminating bias

One of the issues stepping into 2020 has been the susceptibility to bias. For example, small datasets of people may not include all ethnic groups. This can make it impossible for an AI program to recognize or behave appropriately when confronted with these ethnic groups. 

Throughout the year, efforts have been made towards integrating ethnic studies and eliminating bias. However, more actions need to be taken to teach algorithms to behave in a smarter, more natural, and culturally sensitive way for human understanding.

  • Regulation and ethics

Given the rapid adoption of artificial intelligence in high-stakes domains, as well as the concern with potential AI misuse, regulation has long been needed. In fact, in 2020 the European Commission developed a regulatory framework on artificial intelligence that had a wide impact, similar to the GDPR, on any company looking to do business in the EU. 

In the US, legislators are still working on determining a set of policies to regulate the growing industry of artificial intelligence.

Overall, we’re very excited to witness the evolution of AI technology. According to the Gartner report, distributed cloud, AI engineering, cybersecurity mesh and composable business will drive some of the top trends in 2021. More about new emerging technology trends in our next blog post!

So, where will business technology trends in 2021 take us? Find out in our new blog post – Top 5 Business Technology Trends To Watch in 2021