Abstract: Anna Salamon
Can Whole Brain Emulation help us build safe AGI?
An informal poll of the workshop participants at Singularity Summit 2009 found that the vast majority of participants believed that Whole Brain Emulation (WBE) would allow a safer entrance to AGI. We consider three mechanisms whereby WBEs could ease challenges to positive development of further AGI systems. First, existing knowledge of human psychology could be applied to WBE, allowing advance preparation in the selection of candidates for brain emulation and the use of mechanism design to create incentive compatible institutions for WBE. Second, WBEs, as software, could be run at high speeds and with multiple copies, reducing the challenges of monitoring the activities of subsequent, less humanlike, AGI generations. Third, WBEs operating at high subjective speeds would enjoy a much greater abundance of researcher thinking per unit of computational power, favoring substitution of deep theoretical understanding for ‘brute force’ hardware-intensive approaches in subsequent AGI design.