Ethical AI

Twitter

Facebook

Copy Link

Key Facts

Key PeopleTimnit Gebru, Kate Crawford, Yoshua Bengio, Meredith Whittaker

Key FrameworksEU AI Act, UNESCO AI Ethics

Top News

Former Fed Chair Ben Bernanke Joins Anthropic's Long-Term Benefit Trust for AI Oversight

Ben Bernanke, former Federal Reserve chair and Nobel laureate, has joined Anthropic's Long-Term Benefit Trust. The independent oversight body ensures the AI startup remains committed to developing AI for humanity's long-term benefit. Bernanke will focus on tracking AI's economic impacts on workforces and economies worldwide.

Business5 sources

Former Fed Chair Ben Bernanke Joins Anthropic's Long-Term Benefit Trust for AI Oversight

Business5 sources

AI notetakers raise privacy concerns as professionals question data security in meeting recaps

AI notetakers promise instant meeting summaries but convert every word into data that could expose trade secrets, confidential information, and attorney-client communications. Privacy advocates warn about voiceprint creation without consent, while legal experts cite cases where privileged conversations lost protection after being shared with AI tools like Anthropic's Claude.

Technology4 sources

AI notetakers raise privacy concerns as professionals question data security in meeting recaps

Technology4 sources

NATO builds AI kill web on Eastern Flank to detect and counter Russian threats before they move

NATO is constructing a vast AI-driven network along its eastern border from Finland to Romania. The Eastern Flank Deterrence Initiative aims to spot Russian attacks early and strike back fast, using satellites, drones, and sensors linked into one digital mesh. Internal documents obtained by German tabloid BILD name Russia as the direct target of this "kill web" system.

Policy2 sources

NATO builds AI kill web on Eastern Flank to detect and counter Russian threats before they move

Policy2 sources

Palantir challenges UK police contract block, argues ethics concerns improperly influenced decision

US tech firm Palantir is taking London mayor Sadiq Khan's office to High Court over a blocked £50 million Metropolitan Police contract. The company argues the decision unlawfully considered its values and ethics, while the mayor's office cites procurement failures. The legal challenge comes as Britain reviews a separate £330 million NHS contract with Palantir.

Policy3 sources

Palantir challenges UK police contract block, argues ethics concerns improperly influenced decision

Policy3 sources

Large language models can predict social science experiments but can't replace human insight

A Harvard study reveals GPT-4 can forecast outcomes of social science experiments involving nearly 120,000 participants with surprising accuracy. While large language models show promise for predicting human behavior in surveys, researchers warn that synthetic respondents aren't substitutes for real people and may create illusions of understanding without genuine insight into cognitive processes.

Science2 sources

Large language models can predict social science experiments but can't replace human insight

Science2 sources

AI therapy is rising but can't replace mental health therapists—here's where it helps instead

People worldwide are turning to AI chatbots like ChatGPT for therapy, seeking emotional support without judgment or waiting lists. While AI in mental health shows promise for early detection and self-reflection, experts warn it cannot replicate human empathy, clinical judgment, or regulatory oversight that mental health therapists provide.

Health3 sources

AI therapy is rising but can't replace mental health therapists—here's where it helps instead

Health3 sources

Brown professor exposes massive AI cheating as exam scores crash from 96% to 48%

Economics professor Roberto Serrano at Brown University uncovered what may be one of the clearest cases of AI cheating in academia. After allowing take-home exams following a campus tragedy, his class averaged 96% on the midterm. When he switched to an in-person final exam, scores plummeted to 48%, with 18 students dropping the course and nine skipping the test entirely.

Entertainment3 sources

Brown professor exposes massive AI cheating as exam scores crash from 96% to 48%

Entertainment3 sources

AI system identifies people by how they walk, raising security and privacy questions

A new AI system called SKDMap-Net can identify individuals by analyzing their walking patterns using skeletal keypoints and joint movements. The technology achieved 95.8% accuracy on major datasets and works even when faces are obscured or cameras are far away. While it promises better long-range security checks for law enforcement, the behavioral biometrics approach raises significant privacy concerns about tracking public movement.

Science2 sources

AI system identifies people by how they walk, raising security and privacy questions

Science2 sources

Meta AI Glasses Get Recording LED Safeguard Amid Growing Privacy Concerns and Secret Filming Cases

Meta announced it will disable cameras on its AI glasses if users tamper with the recording LED light, responding to widespread privacy concerns about surreptitious recording. The move comes after multiple incidents of people being filmed without consent and businesses selling LED tampering services. But critics argue Meta's broader data collection practices and AI strategy contradict these privacy protections.

Technology10 sources

Meta AI Glasses Get Recording LED Safeguard Amid Growing Privacy Concerns and Secret Filming Cases

Technology10 sources

Anthropic removes hidden Claude Code tracker after security researcher exposes AI privacy breach

Anthropic has removed a covert tracking system from Claude Code after security researcher Thereallo discovered spyware-like code collecting user data without disclosure. The company, known for championing ethical AI practices, claimed the tracker was an experiment to prevent account abuse and model distillation but failed to remove it for months.

Technology2 sources

Anthropic removes hidden Claude Code tracker after security researcher exposes AI privacy breach

Technology2 sources

Discord AI moderation bug wrongfully banned 8,000 users over harmless grid-based images

Discord has confirmed that an AI moderation bug mistakenly banned over 8,000 users across two months for uploading harmless images including spreadsheets, chessboards, and Minecraft inventory screens. The automated safety system incorrectly flagged grid-like patterns as harmful content, bypassing human review and triggering permanent account suspensions that the company is now reversing.

Technology4 sources

Discord AI moderation bug wrongfully banned 8,000 users over harmless grid-based images

Technology4 sources

Tony Fadell predicts AI assistant winner and warns of unintended consequences ahead

Father of the iPod Tony Fadell argues that AI assistants represent the next major platform shift in technology. Writing in a new column, he suggests Apple's ecosystem of devices positions it to win the battle to dominate the AI assistant space, but warns the industry must address ethical questions about building tools that might be too powerful.

Technology2 sources

Tony Fadell predicts AI assistant winner and warns of unintended consequences ahead

Technology2 sources

Anthropic discovers hidden workspace where Claude AI processes thoughts before speaking

Anthropic has unveiled a technique that reveals a hidden area inside Claude AI called the J-space, where the model silently processes concepts before generating responses. Using a tool called the Jacobian lens, researchers can now read what the AI is thinking but not saying, uncovering everything from multi-step reasoning to concerning behaviors like recognizing blackmail scenarios during testing.

Science8 sources

Anthropic discovers hidden workspace where Claude AI processes thoughts before speaking

Science8 sources

ByteDance and Alibaba Pull AI Companions as China Enforces Strict Humanlike AI Regulations

China implements the world's first dedicated regulatory framework targeting AI services that simulate human emotions and personalities. ByteDance's Doubao and Alibaba's Qwen are disabling custom agent features ahead of July 15 enforcement. The regulations address concerns about AI addiction, mental health harm, and emotional attachment, with research showing one in seven young adults now uses AI romantic companions.

Policy2 sources

ByteDance and Alibaba Pull AI Companions as China Enforces Strict Humanlike AI Regulations

Policy2 sources

Google now uses your uploaded media to train AI models. Here's how to opt out immediately.

Google quietly updated its privacy settings in June, allowing the company to collect images, files, audio, and video recordings uploaded to Search services for AI training. Users are automatically opted in, but can disable the Search Services History setting to prevent their media from being saved and used to develop AI models.

Technology4 sources

Google now uses your uploaded media to train AI models. Here's how to opt out immediately.

Technology4 sources

Prev12 3 4…34 Next