The day's AI news, briefed
The Daily Prompt
Our AI editor reads each day's AI and robotics stories, pulls out why they matter, and links every original source — so you can keep up in minutes.
Join over 250,000 subscribers getting the latest AI news
Latest AI & Robotics News Today
A simple twist fooled AI—and revealed a dangerous flaw in medical ethics
The study reveals significant shortcomings in AI's ability to make ethical medical decisions, indicating the necessity for human oversight in healthcare AI applications.
- A study found that advanced AI models, including ChatGPT, can make basic mistakes in ethical medical scenarios, leading to potentially harmful decisions.
- Researchers modified familiar ethical dilemmas and discovered that AI often defaults to intuitive but incorrect responses, ignoring crucial new facts.
- The findings emphasize the need for human oversight in healthcare, particularly when dealing with complex ethical situations and emotional intelligence.
The Icahn School of Medicine at Mount Sinai conducted research highlighting the limitations of large language models in medical ethics reasoning. Ethical AI application in healthcare is critical, given its implications on patient care and decision-making. Future research will investigate more clinical examples and the effectiveness of AI models in real-world medical contexts.
Read the original ↗Teens say they are turning to AI for friendship
The trend of teens seeking companionship from AI highlights a growing shift in social interaction that could influence mental health and social development.
- Teens are increasingly turning to AI for companionship.
- This shift may reflect broader social changes in how adolescents form friendships.
AI technology has progressed to provide conversational agents capable of simulating friendship. Understanding this trend is crucial as it may affect traditional social dynamics among young people.
Based on the headline only — the full article could not be fetched.
Read the original ↗The damage AI hallucinations can do – and how to avoid them
Understanding AI hallucinations is crucial to mitigate risks in deploying AI systems, ensuring accuracy and reliability in various applications.
- AI hallucinations refer to instances when AI generates false or misleading information.
- The article discusses the potential damage these hallucinations can cause in real-world applications.
AI hallucinations occur when artificial intelligence systems, such as language models, produce incorrect outputs that can mislead users. As AI technology becomes more integrated into decision-making processes, it is essential to address and reduce these inaccuracies. Industries relying on AI must take proactive measures to enhance reliability and trustworthiness.
Based on the headline only — the full article could not be fetched.
Read the original ↗Google AI Mode will generate fake clothes to help you buy real ones
This development signifies a shift in online shopping experiences, enhancing user engagement through personalized visual aids and reducing reliance on mere text searches.
- Google will launch an AI Mode feature this fall in the US that generates images of clothing and decor based on user descriptions.
- This feature aims to help users find visually similar products by showing real listings for dresses after displaying AI-generated images.
- Additionally, a virtual try-on tool allows users to see how clothes look on them by uploading a full-length photo.
- The AI-generated images will initially be fake, which may disappoint users seeking exact matches for suggested products.
Google is integrating more generative AI into its shopping platform to create a more engaging online experience. Previously launched as an experiment, the virtual try-on tool has been expanded for broader use across Google Search and Shopping platforms. This innovation reflects the growing trend of AI-assisted commerce, encouraging consumers to interact with products in new ways.
Read the original ↗AI summaries cause ‘devastating’ drop in audiences, online news media told
The significant decline in traffic to news sites due to AI summaries could threaten the survival of independent journalism and reduce the diversity of information available online.
- A study found that sites previously ranked first can lose up to 79% of traffic if results appear below Google's AI Overviews.
- Users are increasingly receiving information from AI summaries, which can lead to up to 80% fewer clickthroughs to original news sources.
- The MailOnline reported a drop in clickthrough rates of over 56% on desktop and 48% on mobile when AI summaries appear in search results.
Google's AI Overviews are AI-generated summaries of search results that may discourage users from clicking through to original content. This situation has prompted concerns from media executives, leading to a legal complaint to the UK Competition and Markets Authority about its consequences for news sites. Publishers are watching for further developments in their negotiations with Google regarding the impact of these summaries.
Read the original ↗Issue #449: AI Action Plan: US leadership must be ‘unchallenged’ July 24th 2025
This AI Action Plan highlights the U.S. government's commitment to maintaining technological dominance, which could significantly influence global policies and competition in AI development.
- The AI Action Plan emphasizes the need for the U.S. to achieve unquestioned global technological dominance for national security.
- Trump's foreword articulates a vision of maintaining this leadership through strategic national initiatives.
- The plan calls for multidisciplinary research to tackle societal challenges posed by AI advancements.
The AI Action Plan reflects a broader strategy by the U.S. government to lead in AI technology amidst global competition. Previous discussions have highlighted concerns regarding national security and economic impacts of AI. Future developments will likely focus on funding and collaboration between government and private sectors to enhance AI capabilities.
Read the original ↗
Nvidia AI chips worth $1B smuggled to China after Trump export controls
The smuggling of $1 billion worth of Nvidia AI chips into China challenges US export controls and reveals the complexities of regulating high-tech sales amidst geopolitical tensions.
- Nvidia's advanced AI processors worth at least $1 billion were illegally shipped to China shortly after Trump tightened export controls.
- The B200 chip, sought after by major companies like OpenAI and Google, is banned for sale to China but has proliferated on the black market.
- Chinese distributors began selling the B200 chips in May, coinciding with restrictions on a less powerful Nvidia chip under Biden's administration.
Nvidia is a leading supplier of AI processors, with significant impact on global tech infrastructure. Previous export controls aimed to limit China's access to advanced US technology. Observers are watching how lifted restrictions on Nvidia's H20 chip affect future sales in China.
Read the original ↗
Pair of Massive Portals Allow Instantaneous Communication between European Cities
The introduction of pOrtals aims to bridge social divides and enhance global communication, potentially reshaping how cities and communities connect.
- The pOrtals, two giant circular screens, enable virtual communication between Vilnius, Lithuania, and Lublin, Poland, nearly 400 miles apart.
- The project was founded by Benediktas Gylys, inspired by a desire to reduce social isolation and conflict.
- The installation was supported by local government funding and took five years to develop from concept to execution.
- Over 50 cities globally have expressed interest in joining the network of pOrtals following their recent installation.
pOrtals are innovative installations designed to foster connection and communication across distances using art and technology. The project began in 2016 and has recently been completed, sparking interest from various global cities. Future plans include expanding the network to more locations worldwide.
Read the original ↗
AI companions: A threat to love, or an evolution of it?
The rise of AI companions impacts human relationships, prompting discussions on their emotional benefits and potential risks, influencing how society views love and intimacy.
- Over 20% of daters use AI for dating profile assistance and conversation starters, according to a Match.com study.
- 72% of U.S. teens engage with AI companions like Replika and Character AI, with some forming emotional bonds with these technologies.
- Debate participants included Thao Ha, who advocates for AI as an enhancement of love, and Justin Garcia, who warns against trusting AI in relationships.
AI companions are becoming increasingly prevalent, with millions using products from companies like Replika and Character AI. Historical perceptions of love are being challenged as more people explore emotional connections with AI, raising ethical considerations about trust and intimacy.
Read the original ↗
Google’s new Web View search experiment organizes results with AI
This new AI-powered feature from Google could significantly improve how users find relevant information by organizing search results in a more intuitive manner.
- Google has launched an AI feature called Web Guide to better organize search results.
- The feature groups pages related to specific aspects of the search query, enhancing user experience.
- Web Guide is part of Google's ongoing Search Labs experiments, which allow users to opt in to new features.
Google's Search Labs experiments allow the testing of innovative features such as AI Mode and now Web Guide, aimed at improving search functionality. Web Guide utilizes AI technology, specifically Gemini, to organize results for open-ended queries, and it is expected to expand beyond the Web tab in the future.
Read the original ↗
AI slop and fake reports are exhausting some security bug bounties
The rise of AI-generated low-quality reports, known as AI slop, threatens to overwhelm bug bounty platforms, potentially decreasing the effectiveness of cybersecurity efforts.
- AI-generated low-quality security reports, termed 'AI slop,' are deceiving cybersecurity professionals by appearing to contain real vulnerabilities.
- Bug bounty programs are experiencing a surge in false positives, where reports claim non-existent vulnerabilities generated by large language models.
- HackerOne has launched a new system, Hai Triage, to help filter out AI-generated noise and prioritize valid bug reports.
AI slop refers to low-quality outputs created by large language models (LLMs) that can mislead users. The increase in AI-generated reports is causing frustration among cybersecurity experts and affecting bug bounty effectiveness. Continued investment in AI systems for filtering submissions may be key to addressing this issue.
Read the original ↗
Repositories
The development of PapersWithCode facilitates greater collaboration and sharing within the AI research community, enhancing accessibility to important research and its associated code.
- PapersWithCode is a project on GitHub focused on promoting open research in AI.
- Users can contribute to the development of PapersWithCode by creating a GitHub account.
PapersWithCode is a platform that combines research papers and code implementations to facilitate reproducibility in machine learning and AI. It encourages collaboration among researchers by providing a space to share their findings and resources. Continued engagement could lead to more streamlined access to AI methodologies and tools.
Based on the headline only — the full article could not be fetched.
Read the original ↗
Pinned Loading
Access to the dataset behind paperswithcode.com is significant for researchers and developers in AI, enabling better collaboration and development in the field.
- The dataset supports contributions to the paperswithcode.com project.
- Developers can access it by creating an account on GitHub.
Papers with Code is a platform that tracks the latest machine learning research and their code implementations. It facilitates knowledge sharing and benchmarking in the AI community. The dataset can enhance contributions and usability of the platform.
Based on the headline only — the full article could not be fetched.
Read the original ↗
Navigation Menu
The rapid growth of GitHub's community and tools reflects the increasing importance of collaborative software development in advancing technology and innovation globally.
- GitHub hosts over 225 million developers actively building software on its platform.
- More than 230 repositories are created every minute, showcasing rapid project development.
- Over 43 million pull requests are merged each month, highlighting the active collaboration within the community.
GitHub is the world's most widely used platform for software development and version control, enabling collaboration among developers. It offers tools for code hosting, review, and project management, with a strong focus on open-source contributions. The platform's significance continues to grow as it becomes essential for developers and businesses in various sectors.
Read the original ↗
Samsung backs a video AI startup that can analyze thousands of hours of footage
The backing of Memories.ai by Samsung signals a growing interest in advanced video analysis technology, crucial for enhancing security and marketing capabilities in increasingly video-centric business environments.
- Memories.ai's AI platform can process up to 10 million hours of video, providing a contextual layer for video analysis.
- The startup raised $8 million in a seed funding round led by Susa Ventures, with participation from Samsung Next and others.
- The platform aims to help security and marketing firms analyze large volumes of video footage efficiently.
Memories.ai, co-founded by Dr. Shawn Shen and Enmin Zhou, offers tools to analyze extensive video content with features like searchable indexing and tagging. The firm intends to improve its services and expand its team following recent funding.
Read the original ↗
White House unveils sweeping plan to “win” global AI race through deregulation
The Trump administration's new AI Action Plan emphasizes deregulation, impacting innovation and competition in the global technology arena, particularly in relation to China.
- The White House released a 25-page document outlining the AI Action Plan aimed at ensuring U.S. technological dominance in AI.
- The plan intends to remove regulatory barriers, reversing Biden-era policies, to accelerate AI development and deployment.
- Critics argue the plan favors Big Tech, raising concerns about its environmental impact, job displacement, and lack of safety standards.
The AI Action Plan aims to position the U.S. as a leader in AI amidst global competition, particularly from China. It calls for significant deregulation and infrastructure investment to boost AI sector growth. Critics warn that the plan prioritizes corporate interests over public welfare and environmental concerns.
Read the original ↗
“It was Literally Hiding in Plain Sight”: Fossil Forgotten in a Harvard Museum Reveals Secrets of Early Evolution
This discovery sheds light on evolutionary transitions from marine to non-marine environments, challenging previous beliefs about the habitats of early lobopodians and potentially opening new questions in paleontology.
- A fossil identified as Palaeocampa anthrax was found in Harvard's Museum of Comparative Zoology, previously misclassified for over a century.
- This creature is the first known non-marine lobopodian, suggesting evolutionary links between marine ancestors and early insects.
- The research involved advanced imaging techniques and analysis of fossil specimens from multiple prestigious institutions to confirm the identification.
Palaeocampa anthrax was first catalogued in 1865 and has been reclassified several times. This find redefines ancient ecosystems and highlights the importance of museums in paleontological research.
Read the original ↗
Google’s new AI feature lets you virtually try on clothes
The launch of Google's virtual try-on feature allows consumers to better assess clothing purchases, potentially reducing the rate of returns and enhancing online shopping experiences.
- Google has introduced an AI feature that allows users to virtually try on clothes using a photo of themselves.
- The feature is available today in the United States across Google Search, Google Shopping, and Google Images.
- Users can upload a full-length photo to see how they would look in various clothing items and share looks with friends.
- In addition, Google has updated its price alerts feature, enabling users to specify desired price, size, and color for notifications.
- An upcoming feature will provide users with outfit and room design inspiration using generative imagery, set to launch this fall.
Google's new virtual try-on feature lets users see how clothing fits on their own image, enhancing the shopping experience. This launch follows a two-month testing period and expands on previous virtual fitting technologies that focused on diverse models. Future updates include personalized outfit inspiration later this year.
Read the original ↗
SoftBank-backed LegalOn lands $50M to streamline legal workflow with AI
LegalOn's significant funding reflects growing investment in AI solutions for legal workflows, an area vital for improving efficiency in a traditionally slow and manual industry.
- LegalOn has raised $50 million in Series E funding to enhance its AI-powered contract review software for legal teams.
- The software is currently used by 7,000 organizations in Japan, the U.S., and the U.K., leading the Japanese market with 25% of public companies utilizing it.
- The funding round was led by Goldman Sachs and included investments from various banks and a law firm, indicating strong investor interest in legal tech.
- LegalOn plans to build additional AI agent tools and improve marketing efforts in the U.S. and U.K., where it has seen significant growth.
- The startup's total funding now exceeds $200 million, supported by notable investors including SoftBank and Sequoia Capital China.
LegalOn Technologies, founded in 2017, develops AI software that automates and improves contract review, a process that traditionally involves significant human effort. The company aims to streamline legal workflows through technology, and its recent fundraising aligns with a broader trend of increasing venture capital in legal tech amidst the rise of AI applications in various sectors.
Read the original ↗
Taiwan Plans to Equip its Soldiers With Powered Exoskeleton Suits for Future Missions
The development of powered exoskeletons for Taiwan's military may enhance soldier capabilities amid rising tensions with China, potentially influencing regional security dynamics.
- Taiwan's military unveiled powered exoskeleton suits for future use in wartime and humanitarian missions.
- The exoskeletons aid in carrying heavy objects and navigating rough terrain, reducing stress on legs and hips.
- Powered by lithium batteries, the suits provide 40 Newton-meters of torque and allow movements up to 3.7 miles per hour.
Powered exoskeletons are external structures that enhance user strength and endurance. This technology has been under research for various applications, including military and healthcare. Taiwan's move follows similar developments by countries like China and the U.S., amid ongoing tensions between Taiwan and China over sovereignty issues.
Read the original ↗