Artificial Speech and Its Authors

摘要: Some of the systems used in natural language generation (NLG), a branch applied computational linguistics, have capacity to create or assemble somewhat original messages adapted new contexts. In this paper, taking Bernard Williams' account assertion by machines as starting point, I argue that NLG meet criteria for being speech actants substantial degree. They are capable authoring messages, and can even simulate illocutionary force speaker meaning. Background intelligence embedded their datasets enhances these capacities. Although there is an open question about who ultimately responsible speech, if anybody, we settle using notion proxy which responsibility artificial acts assigned legally conventionally entity separate from actant.