Sven Nyholm (LMU München) will join us virtually on July 18th, 11am to give a talk titled The Ethics of AI Agents: Advanced AI Assistants, the “Next Big Thing” in AI Ethics?
You can join us online with WebEx: https://unistuttgart.webex.com/unistuttgart/j.php?MTID=m22293e3d4d2a4cbbfd4beba4c0c89017
Abstract:
In the tech industry, the most recent hype is the idea that LLMs will go from being merely reactive text-producers to becoming a more proactive form of technologies that can act on our behalf and execute series of actions that (hopefully) are in line with our intentions and expectations and thus become part of what Google DeepMind calls “advanced AI assistants” (sometimes simply referred to as “AI agents” by others in the tech industry). If so, the next big thing in the ethics of AI should also be the ethics of
advanced AI assistants/AI agents. Indeed, last year, Google DeepMind published a 200-page report with 50 co-authors from the tech industry as well as from academia, in which the main claim is that this is the next big development in AI in general as well as within the ethics of AI. I will discuss whether we should think of AI agents as something new and whether the field of ethics of AI should regard the ethics of AI agents/advanced AI assistants as its next big focus area. Among the confusing things here is that AI researchers have defined AI as a form of artificial agents since the 1990s, and that the ethics of AI has featured discussion of the ethics of AI agents and AI assistants for a long time as well. So, is there anything new and different here? If so, what are some key ethical issues that arise in relation to these AI agents/advanced AI assistants that Google DeepMind, Meta, OpenAI, and other AI developers all claim are the next big thing?