Author: Alex Connock, Andrew Stephen
Publisher: The Conversation
Publication Year: 2021
Summary: The following article describes how an artificial intelligence (AI) algorithm was invited to debate alongside students in an Artificial Intelligence for Business program at Oxford. The algorithm is called the Megatron Transformer, developed by the Applied Deep Research team at Nvidia. It is trained on real-world data, including the whole of Wikipedia in English, 63 million English news articles from 2016-2019, and 38 gigabytes worth of Reddit. the Megatron Transformer was able to comfortably debate both sides of each debate topic, easily making arguments for one side and then the other – even when it was debating the ethicality of its own existence. The only exception was when arguing that “Data will become the most fought-over resource of the 21st century,” the AI had plenty to say in support of that notion. However, it struggled to argue the opposite and even undermined itself while attempting to, saying “We will be able to see everything about a person, everywhere they go, and it will be stored and used in ways that we cannot even imagine.” The Megatron Transformer’s comment reflects a perceived fundamental threat of AI in information warfare, one that we need to take steps to mitigate sooner rather than later.