Chatbots Can Be Trained to Represent Humans after Death: Interview with ChatGPT

Kadir Uludag
DOI: https://doi.org/10.2139/ssrn.4399732
2023-01-01
SSRN Electronic Journal
Abstract:Background: Chatbots such as ChatGPT simulate conversations with human users. Previously, chess bots have trained to represent chess players. Moreover, with advances in language processing, chatbots have become capable of replicating human-like conversations. One area of interest is the skill of training chatbots to represent humans by mimicking their communication. By mimicking them, chatbots can be created to represent a particular person and can be used to reduce people's sorrow after death.Methods: An interview was conducted with ChatGPT to investigate the topic of training chatbots to represent humans. Next, we provided a personal message to allow ChatGPT to respond to the answer accordingly.Results: ChatGPT provided insights into the potential of training chatbots to represent humans. However, it also raised critical ethical considerations like privacy, consent, and deception.Conclusion: ChatGPT provided insights into the potential of training chatbots to represent humans. Evaluating the ethical implications of using personal communication data to train them is vital.
English Else
What problem does this paper attempt to address?