Ever-changing lexicons, multilingualism, and varying cultural value systems compromise accuracy of large language models.
Ever-changing lexicons, multilingualism, and varying cultural value systems compromise accuracy of large language models.
Stanford researchers explore how to build culturally inclusive and equitable AI by offering initial empirical evidence on cultural variations in people’s ideal preferences about AI.
Stanford researchers explore how to build culturally inclusive and equitable AI by offering initial empirical evidence on cultural variations in people’s ideal preferences about AI.
A new white paper for the Congressional Black Caucus addresses risks and opportunities of AI for Black communities.
A new white paper for the Congressional Black Caucus addresses risks and opportunities of AI for Black communities.
Leading policymakers, academics, healthcare providers, AI developers, and patient advocates discuss the path forward for healthcare AI policy at closed-door workshop.
Leading policymakers, academics, healthcare providers, AI developers, and patient advocates discuss the path forward for healthcare AI policy at closed-door workshop.
HAI Associate Director Curt Langlotz explains the current state of health regulation and where we need to move to protect patients and better assist doctors.
HAI Associate Director Curt Langlotz explains the current state of health regulation and where we need to move to protect patients and better assist doctors.
A cross-disciplinary group of Stanford students examines the ethical challenges faced by data workers and the companies that employ them.
A cross-disciplinary group of Stanford students examines the ethical challenges faced by data workers and the companies that employ them.
Scholars develop a new framework that optimizes compound AI systems by backpropagating large language model feedback.
Scholars develop a new framework that optimizes compound AI systems by backpropagating large language model feedback.
The bulk purchase of digital ads through automated placement platforms often matches legitimate ad buyers with websites that traffic in false and misleading information.
The bulk purchase of digital ads through automated placement platforms often matches legitimate ad buyers with websites that traffic in false and misleading information.
A new study adapts large language models to summarize clinical documents, showing a promising path for AI to improve clinical workflows and patient care.
A new study adapts large language models to summarize clinical documents, showing a promising path for AI to improve clinical workflows and patient care.
Digital Economy Lab Fellow Alex “Sandy” Pentland: Building a Better, Safer Digital Ecosystem
The MIT professor describes practical approaches to improving the internet.
The MIT professor describes practical approaches to improving the internet.
In risk modeling, AI researchers take a more-is-better approach to training data, but a new study argues that a less-is-more approach may be preferable.
In risk modeling, AI researchers take a more-is-better approach to training data, but a new study argues that a less-is-more approach may be preferable.
Stanford scholar Tobias Gerstenberg explores how humans judge causation and poses the question: What will happen when AI can do the same?
Stanford scholar Tobias Gerstenberg explores how humans judge causation and poses the question: What will happen when AI can do the same?
A new study reveals the need for benchmarking and public evaluations of AI tools in law.
A new study reveals the need for benchmarking and public evaluations of AI tools in law.
Robotics and hospital computer vision projects receive NSF grants as part of an innovative pilot program to democratize AI research.
Robotics and hospital computer vision projects receive NSF grants as part of an innovative pilot program to democratize AI research.
New studies show a marked spike in LLM usage in academia, especially in computer science. What does this mean for researchers and reviewers?
New studies show a marked spike in LLM usage in academia, especially in computer science. What does this mean for researchers and reviewers?