Next-level AI: does the rise of artificial general and super-human intelligence mark the end of IP? Clark W Lackert World Trade Centers Association 01 June 2024
Image by Microsoft Copilot
Intelligence levels beyond AI as we currently know will require a complete rethink of intellectual property and the IP system, argues Clark W Lackert, deputy general counsel at the World Trade Centers Association, in this week’s opinion column. We therefore need to look beyond ‘narrow AI’ and start discussing the issues of artificial general and superhuman intelligence now, Lackert insists. The 21st Century thus far has presented many challenges to IP rights, from the geopolitical discord making new treaties difficult, to the challenges in creating an IP regime in the legal vacuum of outer space (which I have commented on), to the expanding ease of IP infringement created by technologies such as the Internet, mobile phones and social media. However, the most obvious – yet most ignored – technological challenge to intellectual property now is the exponential growth of AI and the rapidly approaching arrival of human-level computer intelligence called ‘artificial general intelligence’ (AGI), and thereafter super-human intelligence (many times greater than humans) called ‘artificial superintelligence’ (ASI). We need to briefly review the state of play and pose some unanswered inquiries, since we cannot obtain the right answers about AI if we do not ask the right questions. The present-day IP law commentators who have reviewed the current state of AI affairs whereby humans direct AI (socalled ‘narrow AI’, including generative AI) as a tool to undertake certain tasks (eg, creating works of art, searching the Internet for counterfeits and trademark infringements, and devising patentable inventions) try to answer questions such as can software be an author, a trademark owner or an inventor? Issues from narrow AI also arise in areas such as deceiving the public with visual or audio deepfakes – for example, to what level can name, image and likeness be protected? Does data-scraping websites and publications for erudite human-like answers by large language model chatbots constitute fair use? However, little had been written about what is around the next corner – namely, the rapidly approaching age of AGI and thereafter ASI, where software thinks on its own and writes its own code, with no or limited human guidance. Apart from this development, soon humanity will not even understand AI conversations beyond our so-called ‘cognitive horizon’. In these worlds, humans will have little – if any – input into IP development, monetisation, protection and enforcement processes, yet the impact on humans is clear.