In Conversation with AI

Slide 1

• In Conversation with AI
• https://www.uni-heidelberg.de/en/newsroom/in-conversation-how-large-language-models-convincingly-mimic-human-conversation

Slide 2

• Anthropomorphization
• Giving personifying human characteristics to nonhuman entities
• “We personify [language models] output, even though we know it’s just an electrical signal. This is called anthropomorphization.”
• Impact is that it blurs the line between AI and human, which lead to over-trust

Slide 3

• AI “consciousness” through language
• Mimicry of patterns of language
• Development of its own easily-recognizable language patterns
• Personal pronouns
• Humanistic language
• Creative and personal language
• How we talk about AI
• ”Chat with” LLMs
• Ask LLMs for information
• Giving chatbots a name
• Use of em dashes and semicolons
• “It’s not just X, it’s also Y”
• Speech standardization

Slide 4

• Implications
• Subtle shift towards AI-like phrasing and structure
• Reduced linguistic diversity in everyday communication
• Emergence of “AI-register” writing style
• For language
• Perceived authority through conversational fluency
• Anthropomorphism leads to over-trust
• AI framed as agent rather than a tool
• For technology
• For education
• For interaction
• Conversation shifts from tool use to social interaction
• AI treated as participant in communication rather than instrument
• Responsibility becomes diffused across human & machine
• Students self-monitor language to avoid sounding “AI-like”
• Writing becomes performative proof of humanness
• Difficulty distinguishing assistance vs authorship

Slide 5

• Discussion
• 1. Do you think we should make a conscious effort to discontinue the use of humazing language around AI?
• 2. What responsibilities do developers have when designing and implementing conversational AI (human-like responses)?