Latest Articles for Technology Ethics

A framework to understand our reliance on AI in decision-making.

This paper examines prompt injections and their implications for AI models.

Examining how summarization models reflect bias in political opinions.

This study analyzes how LLMs can forecast AI's possible harms.

Examining harm amplification in text-to-image models and its societal impact.

New framework helps generative models forget sensitive data while maintaining performance.

This study investigates jailbreaking attacks on multimodal large language models.

Investigating security risks and detection methods for diffusion models.

Examining how machine learning perpetuates gender biases and their emotional effects.

Examining the relationship between data protection laws and machine learning practices.

Exploring methods to safeguard individual data in an information-driven world.

Exploring how friction can enhance user experiences in AI.

An analysis of the qualities and challenges of language model explanations.

Examining the limitations of LLMs in understanding and retaining temporal information.

New watermarking methods improve text variety and detection in machine-generated content.

An analysis of how attention is captured and its impact on society.

This article examines the dangers of harmful fine-tuning in language models.

New methods secure data in AI while ensuring effective computations.

A method to remove unwanted skills from language models while keeping essential functions intact.

A new benchmark aims to measure and mitigate AI-related dangers.

A framework to evaluate biases in recommendations generated by large language models.

Methods to minimize bias in large language models for fairer outcomes.

This paper analyzes gender bias in large language models and proposes measurement methods.

Evaluating how biases in language models affect real-world applications.

New model creates lifelike images from identity features using machine learning.

Exploring the key traits and challenges of developing reliable AI systems.

A study of techniques used to bypass safety measures in AI language models.

A study that measures political bias in large language models through stance and framing.

A closer look at sparse feature circuits in language models and their implications.

Exploring the importance of understandable reasoning in AI predictions.

A framework to improve the safety and reliability of large language models.

Exploring the role of ethics in language translation technology.

Examining how machine unlearning can expose sensitive data.

ALERT benchmark assesses safety risks in language models to improve their responses.

A new tool for making images safer and more responsible.

Research investigates biases in Turkish language models and their societal impacts.

Learn how Context Steering enhances language model responses through adaptable context use.

Recent research challenges the simplicity of the Knowledge Neuron Thesis in language models.

Research tackles privacy concerns in language models through innovative unlearning methods.

Research reveals bias in AI tools used for hiring based on race and gender.