Flo
@flonarr.bsky.social
@Microsoft pull the plug, hit the big red button, and let's call it a day.
May 5, 2025 at 7:04 AM
@Microsoft pull the plug, hit the big red button, and let's call it a day.
Link zum Projekt www.innerspace.eu
Innerspace - Training & Quality Risk Management
Innerspace helps pharmaceutical companies achieve the highest standards of quality and efficiency in aseptic manufacturing. Training & quality risk management solutions. Contact us to learn more.
www.innerspace.eu
April 1, 2025 at 8:10 AM
Link zum Projekt www.innerspace.eu
Conclusion:
People generally appreciate AI-generated content, but only when unaware of its origin. Once disclosed, personal biases, domain context, and demographic factors influence their preferences.
People generally appreciate AI-generated content, but only when unaware of its origin. Once disclosed, personal biases, domain context, and demographic factors influence their preferences.
March 25, 2025 at 8:57 AM
Conclusion:
People generally appreciate AI-generated content, but only when unaware of its origin. Once disclosed, personal biases, domain context, and demographic factors influence their preferences.
People generally appreciate AI-generated content, but only when unaware of its origin. Once disclosed, personal biases, domain context, and demographic factors influence their preferences.
5. Trust and Perception
• Users associate credibility and authenticity more strongly with human authorship, especially in sensitive areas like health or ethics
• The perception shift underscores how trust in content is shaped by perceived authorship, not just textual quality
• Users associate credibility and authenticity more strongly with human authorship, especially in sensitive areas like health or ethics
• The perception shift underscores how trust in content is shaped by perceived authorship, not just textual quality
March 25, 2025 at 8:57 AM
5. Trust and Perception
• Users associate credibility and authenticity more strongly with human authorship, especially in sensitive areas like health or ethics
• The perception shift underscores how trust in content is shaped by perceived authorship, not just textual quality
• Users associate credibility and authenticity more strongly with human authorship, especially in sensitive areas like health or ethics
• The perception shift underscores how trust in content is shaped by perceived authorship, not just textual quality
4. Context Is Key
• Preference for human vs. AI content varies across domains:
- Humanities and Life Sciences: Greater bias toward human authorship when the source is revealed.
- Physical and Social Sciences: Less impact of disclosure on preference.
• Preference for human vs. AI content varies across domains:
- Humanities and Life Sciences: Greater bias toward human authorship when the source is revealed.
- Physical and Social Sciences: Less impact of disclosure on preference.
March 25, 2025 at 8:57 AM
4. Context Is Key
• Preference for human vs. AI content varies across domains:
- Humanities and Life Sciences: Greater bias toward human authorship when the source is revealed.
- Physical and Social Sciences: Less impact of disclosure on preference.
• Preference for human vs. AI content varies across domains:
- Humanities and Life Sciences: Greater bias toward human authorship when the source is revealed.
- Physical and Social Sciences: Less impact of disclosure on preference.
3. Gender & Expertise matter
• women prefer human responses, regardless of source.
• men show bias only if AI origin is disclosed.
• users with programming skills favor human responses when aware it’s AI, likely due to awareness of LLMs’ limitations.
• women prefer human responses, regardless of source.
• men show bias only if AI origin is disclosed.
• users with programming skills favor human responses when aware it’s AI, likely due to awareness of LLMs’ limitations.
March 25, 2025 at 8:57 AM
3. Gender & Expertise matter
• women prefer human responses, regardless of source.
• men show bias only if AI origin is disclosed.
• users with programming skills favor human responses when aware it’s AI, likely due to awareness of LLMs’ limitations.
• women prefer human responses, regardless of source.
• men show bias only if AI origin is disclosed.
• users with programming skills favor human responses when aware it’s AI, likely due to awareness of LLMs’ limitations.
2. Bias Against AI—When Revealed
• Once users are told that a response is AI-generated, their preference shifts significantly towards human content.
• This demonstrates a bias against AI, driven more by the label than by actual content quality.
• Once users are told that a response is AI-generated, their preference shifts significantly towards human content.
• This demonstrates a bias against AI, driven more by the label than by actual content quality.
March 25, 2025 at 8:57 AM
2. Bias Against AI—When Revealed
• Once users are told that a response is AI-generated, their preference shifts significantly towards human content.
• This demonstrates a bias against AI, driven more by the label than by actual content quality.
• Once users are told that a response is AI-generated, their preference shifts significantly towards human content.
• This demonstrates a bias against AI, driven more by the label than by actual content quality.
1. Quality Over Source—At First
• When users don't know the origin of a response, they tend to prefer LLM-generated answers over human-written ones.
• AI responses are often longer, more structured, and more informative, which likely contributes to their appeal.
• When users don't know the origin of a response, they tend to prefer LLM-generated answers over human-written ones.
• AI responses are often longer, more structured, and more informative, which likely contributes to their appeal.
March 25, 2025 at 8:57 AM
1. Quality Over Source—At First
• When users don't know the origin of a response, they tend to prefer LLM-generated answers over human-written ones.
• AI responses are often longer, more structured, and more informative, which likely contributes to their appeal.
• When users don't know the origin of a response, they tend to prefer LLM-generated answers over human-written ones.
• AI responses are often longer, more structured, and more informative, which likely contributes to their appeal.