Explainable AI (XAI) is a powerful concept that comes from questioning the reliability of artificial intelligence (AI). It refers to the ability to explain the decisions, recommendations, predictions, and other similar actions made by an AI system. Recent years have seen a rapid expansion of AI implementation, but as models become more specialised and complex, they turn out to be increasingly opaque and harder to interpret. In short, the wide-scale commercialization and spread of AI hasn’t gone hand-in-hand with the development of control mechanisms and transparency efforts.

This is why companies are now relying on XAI to bring the additional layer of interpretability, especially in such highly complex processes like AI surgery or autonomous driving that require the ability to explain and trace back the decision-making in case of an accident. Likewise, social media platforms like Facebook are using AI to tag specific news stories as fake – but it may be useful to see why. The same applies to decisions about loans, insurance claims, or legal contract analyses, where explainability becomes almost mandatory from both regulatory and business practice standpoint.

Giants like Google and IBM are already making the best out of XAI: The latter has developed a cloud-based AI tool that empowers users by showing them which major factors determined the AI-based recommendation, uncovering issues like inherent bias in real-time. The use of XAI also goes well beyond the private sector. For example, the US Department of Defense has championed international efforts to make AI models easier to pick and understand.

With XAI becoming more important, we decided to take a closer look at some of the most highly ranked companies we track in this space.

Founded by academics at the University of Waterloo in Canada, DarwinAI uses AI to obtain a foundational understanding of how a neural network based on deep learning works and then applies that understanding to generate highly compact versions of that model. The process, known as “Generative synthesis,” enables developers to understand, interpret, and defend the inner workings of their models and how they reach decisions. With its explainability toolkit, the company offers a simple-to-use feature for network performance diagnostics, particularly helpful for network debugging, design improvement, and addressing regulatory compliance.

Flowcast uses patented machine learning technology to create predictive credit-risk models that enable lenders to unlock capital for small to medium businesses. As a credit decision requires both an accurate assessment of risk and a simple explanation, Flowcast developed systems that allow an accompanying plain English explanation. The resulting statements, such as “Client A is rejected because their months since most recent diluted payment is 2 (1.8 above median), and the USD amount requested is $72K ($57K above median)”, are used to build trust with Flowcast’s banking partners’ credit risk officers.

Imandra democratises advances in automated reasoning to make algorithms safe, explainable and fair. Its “Reasoning as a Service” feature brings these techniques closer to those without a specialised background in these fields. Starting off in the financial sector, Imandra has worked with major investment players, who benefit from designing, testing, and conducting audits of their complex trading systems. But with constant growth, the company’s client base has expanded into robotics, autonomous vehicles, and reinforcement learning.

Kyndi, an explainable natural language processing platform, is on a mission to optimise human cognitive performance by transforming business processes through auditable AI systems. It uses an established programming language called Prolog to automate fact, inference, and concept generation in almost any vertical. Its elaborate model helps encode the meaning of documents, repositories, and domains in queryable knowledge graphs. This powerful system has recently earned the company a $20 million investment to further expand its engineering and sales capabilities.

Factmata, established in 2017, was one of the first global companies to tackle fake news online. Its goal is to build a fair, explainable, and open approach to rating content online, rather than judging low credibility based on the alignment of the content with our own opinions. “We believe there are reasonable bounded indicators of good quality, balanced, trustworthy journalism. With enough time and training data, a well-built AI should be able to automatically detect writing that strays from these bounds, whilst leaving the final evaluation and critical opinion to the reader,” says Co-Founder and CEO, Dhruv Ghulati. Factmata’s dedication has worked to attract investors: eyeo GmbH, the company behind Adblock Plus, announced it will support Factmata to develop an online fact-checking browser extension.

Logical Glue is helping businesses unlock the technological power of AI and machine learning. Among other things, it does so by boosting predictive intelligence, enabling data analysis, and discovering trends and patterns in processes. Recently, it was acquired by the Swiss banking software provider Temenos, mainly aiming to advance manual underwriting through AI automated decision-making and recommendations. According to Temenos, the Logical Glue’s Explainable AI platform addresses one of the key challenges for banks using AI: the lack of insights into the automation decision-making.

Deep learning tools have been developed to simulate operations of the human brain. But there’s no wonder that the “black box” nature of such models has led to urgent calls for more transparency across industries. Various companies have embarked on this journey, working to stabilise our confidence in AI systems and ensuring that models are complaint to regulations, avoid biases, and guide better decisions and safer future overall.

If you want to find out more about the companies working on Explainable AI, go ahead and search our database. You can also explore other industries there too, or alternatively you may like to find out about our company scouting services.