A Conversational Brain-Artificial Intelligence Interface

Anja Meunier,Michal Robert Žák,Lucas Munz,Sofiya Garkot,Manuel Eder,Jiachen Xu,Moritz Grosse-Wentrup

2024-03-15

Abstract:We introduce Brain-Artificial Intelligence Interfaces (BAIs) as a new class of Brain-Computer Interfaces (BCIs). Unlike conventional BCIs, which rely on intact cognitive capabilities, BAIs leverage the power of artificial intelligence to replace parts of the neuro-cognitive processing pipeline. BAIs allow users to accomplish complex tasks by providing high-level intentions, while a pre-trained AI agent determines low-level details. This approach enlarges the target audience of BCIs to individuals with cognitive impairments, a population often excluded from the benefits of conventional BCIs. We present the general concept of BAIs and illustrate the potential of this new approach with a Conversational BAI based on EEG. In particular, we show in an experiment with simulated phone conversations that the Conversational BAI enables complex communication without the need to generate language. Our work thus demonstrates, for the first time, the ability of a speech neuroprosthesis to enable fluent communication in realistic scenarios with non-invasive technologies.

Human-Computer Interaction,Artificial Intelligence,Signal Processing

What problem does this paper attempt to address?

This paper introduces a novel Brain-Artificial Intelligence Interface (BAI) that differs from traditional Brain-Computer Interfaces (BCIs) by utilizing the power of artificial intelligence to replace certain neural cognitive processing. While traditional BCIs assume users have complete cognitive abilities, BAIs allow users to accomplish complex tasks by providing high-level intents, with the low-level details determined by pre-trained artificial intelligence agents. This approach expands the target user population of BCIs to include individuals with cognitive impairments who typically cannot benefit from traditional BCIs.

The paper specifically mentions an EEG-based conversational BAI example, which demonstrates the ability to engage in complex communication without generating language, as shown in simulated telephone conversations. By decoding users' intentions from their neurophysiological data and inputting these intentions into an AI system, the AI system generates specific responses. The experiments indicate that this speech neuroprosthesis can facilitate smooth, non-intrusive communication in real-life scenarios.

The paper explores the basic structure, information flow, and main components of BAIs, including context input, cognitive probing, intent decoding, and action generation. Additionally, it discusses the potential applications of conversational BAIs in individuals with language impairments, particularly for individuals with aphasia who can utilize BAIs for language communication. The paper also proposes a concept verification of an EEG-based conversational BAI system (EEGChat), which relies on non-invasive EEG recordings and an automated dialogue agent that allows users to communicate by selecting pre-generated keywords.

Lastly, the paper reports experimental results demonstrating the effectiveness of EEGChat in simulating dialogue tasks and discusses the potential of BAIs in expanding human capabilities while addressing ethical concerns and future research directions.