Dr. Xi Nan, from the University of Sydney in Australia, delivered a keynote speech titled Detection of Tax Evasion Activities on Social Media Platforms: Insights from an Experiment on Instagram and Its Regtech Tool. She concluded that social media has a growing impact on e-commerce, and many transactions are made on social media. Social medium plays a big role in our life, but as a private sharing platform, social media platform transactions are entirely out of the administration by tax authorities, which is considered as “hidden transactions”. Through the convenience of social media platforms as a means of communication, content sharing and even payment between different users, black market transactions are now enabled by e-commerce. Tax authorities have the difficulty to detect such transactions. Since technological innovation has brought in a digitalized version of the hidden economy, when considering combating these tax evasion issues, technological solutions are required to assist the application of existing anti-evasion measures. So Dr. Xi Nan presented a machine learning based Regtech tool for international tax authorities to detect transaction-based tax evasion activities across social e-commerce.

       Based on the dataset, her team developed a multi-modal deep neural network to automatically detect suspicious posts. The proposed model combines comments, hashtags and image modalities (including extracted from videos) to produce the final output. It could help tax authorities to identify audit target in an efficient and effective manner. Based on these data, her team found that When the post is related to the black-market sale and will result in tax evasion, it is observed that the more developed countries are, the less they engage in business activities on social media platforms. When the post is related to the black-market sale and will result in tax evasion, it is examined that 75.2% of the posts will leave contact information of other social software accounts. Such as WhatsApp, Line, WeChat. By using the demo results of the Regtech tool, it gives the detection score of an Instagram post to signal its relevance to transaction-based tax evasion activities, considering the comments, hashtags, images, and their combinations within the post. The detection score ranges from 0 to 1, where “1” means the machine regards the post as highly suspicious black-market sales.