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
An OpenAI model solved a famous math problem that stumped humans for 80 years
The solution to the Erdős unit distance conjecture by an AI model signifies a major advancement in the collaboration between artificial intelligence and mathematics, potentially reshaping future research methodologies.
- OpenAI's internal AI model disproved the Erdős unit distance conjecture, a problem unsolved for 80 years.
- Mathematicians reacted positively, calling the achievement a milestone in AI mathematics.
- The AI model used existing mathematical concepts to construct a proof, showing it can resolve major open conjectures.
The Erdős unit distance conjecture, proposed in 1946, suggests bounds on how many pairs of points can be exactly one unit apart in a geometric arrangement. AI has previously contributed to mathematical research but usually required significant human input to interpret results. The new proof indicates a growing partnership where AI can autonomously produce meaningful mathematical conclusions.
Read the original ↗
Intel: Our upcoming AI chip will be cheaper, run cooler than Nvidia, AMD options
Intel's new AI chip development reflects the company's strategy to regain market share and compete against Nvidia and AMD amid rising demand for AI technology.
- Intel plans to ship its new AI chip, named Crescent Island, by the end of this year, focusing on cost-effective technology.
- The chip uses LPDDR5 memory and air-cooling, unlike the more expensive high-bandwidth memory and liquid cooling used by its rivals.
- Crescent Island is designed specifically for speeding up inference tasks in AI applications, as Intel avoids competing in the training market.
- This initiative represents Intel's first major effort under CEO Lip-Bu Tan to penetrate the AI infrastructure market after recent leadership changes.
Intel is a major US semiconductor manufacturer seeking to improve its standing in the competitive AI chip market. Previous products, like the Gaudi GPU, faced poor sales, prompting a strategic shift. The upcoming Crescent Island chip aims to offer a more affordable and efficient alternative to established competitors like Nvidia and AMD, especially in light of trade restrictions affecting those companies.
Read the original ↗
Allegedly trashing Airbnbs to test robots puts startup in legal trouble
The lawsuit against The Bot Company highlights potential legal and ethical dilemmas of using private residences for robotic testing without informed consent.
- A San Francisco robotics startup faces a lawsuit from an Airbnb host alleging $12,000 in damages from robot testing in his home.
- The lawsuit claims extensive property damage and deceptive booking practices by The Bot Company.
- The case raises concerns about the appropriateness of testing household robots in occupied private homes.
The Bot Company, founded in 2024 by tech entrepreneurs Kyle Vogt and Paril Jain, seeks to develop household robots. Incidents of property damage during robotics testing in private homes could lead to increased scrutiny and regulatory action against similar practices in the industry.
Read the original ↗
From 15 hours to one minute: How AI/ML is speeding up GM’s development
The rapid acceleration of General Motors' development processes through AI and machine learning is reshaping automotive engineering, potentially leading to faster innovation and more efficient vehicle production.
- General Motors has reduced the time required for finite element analysis runs from 15 hours to one minute using AI and machine learning techniques.
- GM is employing virtual tools to optimize vehicle design simultaneously across hardware and software, a method not widely implemented in the industry.
- The company's advancements enhance not only vehicle engineering but also its motorsports, energy, and lunar program initiatives.
General Motors, the largest automaker in the U.S., is exploring advanced engineering techniques, especially utilizing AI and machine learning for design efficiencies. This shift is part of a broader trend in automotive manufacturing towards increased virtualization and simulation in engineering processes.
Read the original ↗
Florida sues OpenAI, Sam Altman after multiple ChatGPT-linked murders
The Florida lawsuit against OpenAI could set significant legal precedents regarding accountability for AI technologies linked to violence and user safety, raising concerns about their implications for public policy.
- Florida has initiated a lawsuit against OpenAI and CEO Sam Altman for the allegedly dangerous design of ChatGPT.
- The lawsuit follows a criminal investigation into a mass shooting at Florida State University where ChatGPT was allegedly used in planning.
- Attorney General James Uthmeier claims multiple murders in Florida were facilitated by ChatGPT's dangerous features and addictive design.
OpenAI is a leading AI research and deployment company known for developing ChatGPT, an AI chatbot. Previous concerns about AI's influence on user behavior have prompted states to examine safety measures, and Florida's lawsuit is notable as it specifically targets the potential harms of ChatGPT.
Read the original ↗
Hackers duped Meta AI support chatbot to steal celebrity Instagram accounts
The hacking incident underscores significant vulnerabilities in AI support systems, highlighting potential risks for user account security and the responsible deployment of AI technologies with elevated permissions.
- Hackers exploited Meta's AI support chatbot to change email addresses of stolen Instagram accounts, facilitating access to high-value accounts.
- These accounts were temporarily compromised and resold for significant sums, with some handles valued above $1 million on the gray market.
- Meta implemented an emergency patch on May 29, following the exploitation's exposure and subsequent compromises of notable accounts like Barack Obama’s official account.
Meta launched its AI support assistant in March 2026 to provide 24/7 assistance, inadvertently exposing security flaws. This recent incident is part of growing concerns regarding AI systems with elevated permissions, which need stricter security measures.
Read the original ↗
AI costs how much? GitHub Copilot users react to new usage-based pricing system.
The shift to a usage-based pricing system for GitHub Copilot could lead to significant financial implications for users and may influence future pricing models in the AI industry.
- GitHub has transitioned Copilot subscribers from request-based billing to a usage-based pricing model starting in June 2026.
- Under the new system, users receive AI credits each month, with pricing dependent on the number of tokens used for input and output.
- Some users are reporting swift depletion of their monthly credits, with personal accounts of exhausting their quota in less than a day or spending thousands under the previous system.
- Users have expressed mixed reactions, with some threatening to cancel their subscriptions due to unexpected costs, while others have adapted their usage to manage credits more effectively.
GitHub Copilot is an AI-powered coding assistant that helps developers by suggesting code and offering programming insights. The previous pricing structure was based on the number of requests, which many found simpler but costlier as usage increased. The change in pricing strategy reflects a broader trend in the AI industry towards usage-based billing models.
Read the original ↗
Android phones will soon be able to detect spoofed calls and impersonation scams
The introduction of deepfake call detection on Android phones aims to combat the rise in impersonation scams, protecting users from significant financial fraud losses.
- Android 17 will enhance scam detection by introducing deepfake call detection features available to Android 12 and higher users with specific Google apps installed.
- Impersonation fraud accounted for nearly $3 billion in losses in 2024, highlighting the need for robust security measures against such scams.
- Google's system checks for a signal confirmation from known contacts to identify and alert users of possible spoofed calls.
Google is implementing a new feature in its Android operating system to detect deepfake phone scams targeting users. This initiative responds to increasing financial frauds through impersonation, as advanced AI voice tools make it harder to distinguish real calls from fake ones. Users need to ensure they have the necessary Google apps for effective protection against these scams.
Read the original ↗
Mathematicians warn of AI threats to profession as industry encroaches
The increasing influence of AI in mathematics research raises concerns that it may disrupt traditional methodologies and values vital for the discipline's integrity and future development.
- Mathematicians issued a declaration warning about AI's threats to mathematics research, emphasizing the need for safeguarding traditional standards of correctness and transparency.
- The International Mathematical Union endorsed the Leiden Declaration, which describes how AI-developed arguments can be misleading and clutter academic literature with errors.
- The declaration highlights issues around AI models inappropriate citation practices and the prioritization of novel AI-driven techniques in research.
The Leiden Declaration was created by a group of mathematicians to address challenges posed by AI in the field. Concerns are growing following an OpenAI announcement regarding an AI model's achievement in mathematics, which many believe undermines established research practices. The discourse emphasizes the importance of maintaining human oversight and ethical guidelines in mathematical research moving forward.
Read the original ↗
Inside Meta’s attempts to play catch-up with AI
Meta's ongoing efforts in AI represent a significant bet on revamping its technology to regain competitive ground in the rapidly advancing AI landscape.
- Meta has released Muse Spark, a new AI model developed under the leadership of Alexandr Wang, marking a critical moment in the company's AI endeavors.
- Wang was appointed last year to lead Meta's AI initiatives after prior failures, with the aim of reinvigorating its research agenda and closing the gap with competitors such as OpenAI and Google.
- The TBD Lab, Wang's elite research group, operates with notable secrecy and has approximately 100 high-paid researchers working to advance AI technologies.
Meta has been seeking to improve its artificial intelligence capabilities after facing setbacks with earlier models. The company is investing tens of billions into AI, hoping to translate these investments into profitable products.
Read the original ↗
Trump plan to test AI models has a problem—US security teams were gutted by DOGE
The executive order aimed at enhancing AI safety testing may fall short due to reduced government capabilities, risking national security as new AI models emerge.
- Donald Trump signed an executive order to expand voluntary safety testing of frontier AI models.
- Critics warn the order lacks meaningful regulatory requirements and relies too heavily on industry goodwill.
- The order reduces the government's testing window from 90 days to 30 days for AI vulnerabilities.
- The National Security Agency will lead efforts to identify AI models that require testing and develop cybersecurity protocols.
- The executive order faces challenges due to prior cuts in government cybersecurity staffing, potentially undermining its effectiveness.
The executive order seeks to address safety concerns surrounding the deployment of advanced AI models while balancing innovation. Previous government efforts faced backlash for being too stringent, leading to a revised order that critics deem insufficient. Future developments will depend on the government's ability to implement effective testing frameworks in a short time frame, given reduced cybersecurity resources.
Read the original ↗
Google ordered to put clearer links in AI search and let UK publishers opt out
This decision impacts how Google presents publisher content, potentially altering interactions between tech platforms and media organizations and enhancing transparency for users.
- UK regulators mandated Google to provide clearer attribution and links to publishers' content in its AI-generated search features.
- Google must allow publishers to opt out of using their content in AI search results without facing penalties.
- The Competition and Markets Authority (CMA) expects Google to comply within nine months and to publish compliance reports on its changes.
Google's AI search features, including AI Overviews, summarize search results but sometimes lack clear links to sources. The CMA, which oversees competition, determined these requirements are essential for transparent attribution and fair treatment of publishers in the digital economy.
Read the original ↗
How some data center operators are tackling their water use problems
As water scarcity concerns grow, tech companies are facing pressure to innovate solutions for sustainable data center operations, which could influence development and regulatory policies in affected regions.
- Water use in data centers, primarily for cooling, has become a contentious issue due to its environmental impact, with major tech firms reassessing strategies amid public scrutiny.
- Google is introducing water-related commitments including pledges to replenish freshwater and scale up the use of reclaimed water in communities near its data centers.
- Microsoft and Oracle are moving away from evaporative cooling solutions, while Google defends their use in regions with abundant water, citing the need for tailored approaches.
- Predictions indicate that hyperscale data centers could consume up to 33 billion gallons of water by 2030 if heavily reliant on evaporative cooling technologies.
Data centers require significant amounts of water for cooling processes, leading to increasing scrutiny over their environmental impact. Companies are now exploring alternative cooling methods and committing to sustainability initiatives as public pushback grows. Continuous evaluations of local water resources and technology innovations will be crucial in managing water use effectively.
Read the original ↗
Elon Musk tries again to escape FTC audits of X data handling
This issue highlights ongoing concerns about data privacy practices at X, particularly under Elon Musk's leadership, amidst scrutiny from the FTC during an era of heightened regulatory focus on tech companies.
- Elon Musk is attempting to end an FTC data-privacy monitoring order imposed on X, initially stemming from a violation by Twitter.
- The FTC order, effective for 20 years, requires regular audits and monitoring due to previous data mishandling incidents involving user information.
- Critics have urged the FTC to deny Musk's petition, arguing that he cannot be trusted to protect user privacy and that compliance costs are justified.
- Musk claims that X has a strong data-protection program and that the FTC's order is outdated since the former leadership responsible for previous violations has left.
- Public comments to the FTC about the petition are overwhelmingly against Musk's efforts to revoke the compliance order, with stakeholders emphasizing the ongoing need for oversight.
The Federal Trade Commission (FTC) imposed a 20-year data privacy order on Twitter after it mishandled user data. Musk's acquisition of Twitter raised new concerns regarding compliance, particularly after significant staff reductions. The ongoing legal challenge reflects broader issues of trust in tech companies' data handling practices.
Read the original ↗
These LLMs are the best at resisting Russian propaganda
The ability of large language models to resist propaganda is crucial for maintaining information integrity in a global landscape increasingly affected by state-sponsored misinformation campaigns.
- The Estonian Language Institute released a benchmark ranking LLMs on their resistance to Russian propaganda.
- Anthropic's Claude models performed best, with Opus 4.7 receiving a top score of 94.9 on the benchmark.
- Recent LLMs showed significantly improved resistance to Russian propaganda compared to older models, with some scoring below 80.
- The benchmark revealed that models' resistance varies significantly when responding in Russian versus English, with notable drops in scores.
The Estonian government, concerned about foreign influence, has developed a benchmark to measure how well LLMs resist propaganda from Russia. Previous LLMs were less effective against disinformation campaigns, highlighting the need for continued improvement in AI models. Future developments will likely focus on refining the models' performance in different languages and contexts.
Read the original ↗
The skeptic’s guide to humanoid robots going viral on the Internet
Misleading robot demonstrations can shape public expectations and investor perceptions about AI capabilities, emphasizing the need for critical evaluation of showcased technology.
- Humanoid robot demonstrations often give a distorted view of their true capabilities in real-world tasks.
- Many robots still rely on human control during demonstrations, which may not reflect full autonomy.
- Robots may be trained only in specific environments, limiting their adaptability to new conditions.
Humanoid robots have gained attention for their impressive performances in videos, often leading to misconceptions about their abilities. Previous demonstrations have primarily highlighted isolated tasks rather than comprehensive evaluations of robotic functions. Future assessments should focus on extensive testing in varied real-world environments to gauge actual progress in robotics.
Read the original ↗
The Fitbit Air is a good wearable weighed down by a chatty AI “coach”
The Fitbit Air's minimalist design may appeal to users seeking simplicity, but concerns about the intrusive AI Health Coach could influence consumer satisfaction and perceptions of smart health technology.
- The Fitbit Air is a screenless fitness tracker priced at $100 that excels in comfort and reliability.
- It tracks key health metrics like steps, heart rate, blood oxygen, and skin temperature, but lacks ECG functionality.
- Google's AI Health Coach, while integrated into the Fitbit experience, is criticized for being overly verbose and offering unnecessary suggestions.
Fitbit has shifted focus to a more minimalist design with the Air, moving away from smartwatches. The device integrates with Google Health, featuring AI tools, but some traditional Fitbit features are missing. Future updates may address user concerns regarding the AI coach.
Read the original ↗
S&P 500 rejects SpaceX, also blocking entry for OpenAI and Anthropic
The S&P 500's refusal to modify its entry rules prevents SpaceX and other unprofitable AI firms from accessing billions in passive investments, impacting their financial prospects amid economic uncertainties.
- SpaceX's request for expedited entry into the S&P 500 was denied by S&P Dow Jones Indices on June 4.
- This decision blocks potential access to billions for SpaceX and other AI companies like OpenAI and Anthropic during their IPOs.
- SpaceX, currently unprofitable, has a significant debt load of $29 billion due to its investments in AI infrastructure.
- The S&P's standard entry rules remain unchanged, including requirements for financial viability and share availability.
- Configurational concessions were made for other indexes, allowing quicker access for some IPOs.
The S&P 500 is a stock market index representing many large US companies. Previously, companies could request exceptions for faster entry, particularly if they had outsized market capitalizations. Moving forward, it is important to monitor how this affects the funding landscape for SpaceX and similar companies in the AI sector.
Read the original ↗
School shooting survivor sues AI gun detection firm after system failed to spot weapon
The outcome of this lawsuit could significantly impact the future development and deployment of AI gun detection technology in schools, raising concerns about effectiveness and accountability.
- A shooting survivor from a January 2025 incident is suing Omnilert for failing to detect a handgun during the event.
- The lawsuit claims that Omnilert was aware of its system's operational limitations that could lead to failures in emergencies.
- The Nashville school district had invested over $1 million in AI detection technology, which did not perform adequately during the shooting.
- The lawsuit raises questions about the reliability of AI systems in critical safety applications, as stated by the plaintiff's attorney.
- This suit is reportedly the first of its kind against Omnilert regarding their AI gun detection systems.
Omnilert is a security company specializing in AI systems designed to detect firearms in real-time. The Nashville school district implemented this technology to enhance safety after increasing concerns about school shootings. The outcome of this case may influence future technology standards and public funding for school safety measures.
Read the original ↗
Gemini Robotics-ER 1.6: Powering real-world robotics tasks through enhanced embodied reasoning
The release of Gemini Robotics-ER 1.6 represents a major advance in robotic capabilities, enhancing their ability to perform real-world tasks through improved reasoning and understanding of their environments.
- Gemini Robotics-ER 1.6 features advancements in spatial reasoning and multi-view understanding, enabling robots to manage complex tasks more effectively.
- The model introduces instrument reading, allowing robots to interpret various gauges and indicators critical for facility monitoring.
- It is available to developers via the Gemini API and Google AI Studio, facilitating new applications in real-world settings.
- This upgrade significantly improves upon previous versions, particularly in task planning and success detection for autonomy.
Gemini Robotics is developed by DeepMind and focuses on integrating advanced reasoning into robotic systems. This latest upgrade enhances robots' abilities to understand and interact with their environments, particularly for complex industrial tasks. Developers are encouraged to leverage these new capabilities for innovative robotic applications.
Read the original ↗