Artificial Intelligence News Live: Embedding audit trails in AI systems before scaling deemed crucial for enterprises

The growing deployment of AI applications and agents necessitates that enterprises establish robust and auditable AI pipelines, with a critical emphasis on embedding audit trails into these systems before they are scaled, VentureBeat says in a report. This proactive approach is essential because retrofitting comprehensive logging—which includes data lineage, model parameters, inputs, outputs, and human interventions—into complex, widely deployed AI systems is significantly more challenging and costly than designing for auditability from the outset. The push for such measures is reinforced by regulatory developments like the EU AI Act and frameworks such as the NIST AI Risk Management Framework, alongside an increased industry focus on AI governance and Responsible AI practices, ensuring that AI processes and decisions are traceable and risks can be effectively managed.

New York state lawmakers passed a bill on Thursday, June 13, 2025, known as the RAISE Act, aimed at preventing highly capable "frontier" artificial intelligence models from contributing to large-scale disasters, TechCrunch said in a report. The legislation targets AI systems, such as those developed by OpenAI, Google, and Anthropic, that could potentially lead to scenarios involving the death or injury of over 100 people or causing more than $1 billion in damages. The passage of this bill through the state legislature, which now awaits the governor's signature to become law, represents a significant development for proponents of AI safety and regulation.

Political consultant Steven Kramer was acquitted by a New Hampshire jury on Friday of voter suppression and impersonating a candidate charges related to AI-generated robocalls mimicking former President Joe Biden, AP said in a report. Kramer had admitted to orchestrating the message, which used an AI-generated voice similar to Biden's and his catchphrase "What a bunch of malarkey," sent to thousands of New Hampshire Democrats two days before the state's primary. Prosecutors alleged the calls suggested that voting in the primary was not necessary.

OpenAI has stated its intention to continue working with Scale AI, a key data annotation provider, despite a recent multibillion-dollar investment in Scale AI by Meta Platforms Inc., Bloomberg says in a report. OpenAI's Chief Financial Officer confirmed this decision, emphasizing the company's strategy to maintain access to a broad ecosystem of data sources crucial for developing its artificial intelligence models. This development underscores the importance of data supply chain resilience for AI companies.

Chinese authorities have released draft guidelines concerning the overseas transmission of data collected by automobiles, a development that could pave the way for Tesla Inc. to expand its advanced AI-driven driver-assistance systems, specifically its Full Self-Driving (FSD) software, in the significant Chinese market, Bloomberg said in a report. These proposed rules address previous strict regulations that limited vehicle data export, a challenge for companies like Tesla that utilize such data for global AI model training. While the guidelines are currently in draft form and subject to public comment and potential changes before becoming final, their potential finalization could remove a key regulatory obstacle for Tesla's AI development and deployment in China.

Abu Dhabi’s prominent state-backed artificial intelligence and technology holding group, G42, is establishing a European unit, marking a significant step in the emirate's wider push into Europe, Bloomberg says in a report. This expansion aligns with Abu Dhabi's broader strategy, which includes pledges of billions of dollars in investment by the UAE government across various European sectors, aimed at fostering technological and economic ties. G42's move signals the growing international ambition of the UAE in the field of artificial intelligence.

A team of four-legged robots from Boston Dynamics, identified as likely being their Spot model, executed a synchronized dance routine to Queen's "Don't Stop Me Now" during an appearance on "America's Got Talent," TechCrunch said in a report. This pre-arranged segment, framed as an audition, served to demonstrate the robots' sophisticated control, stability, and capacity for complex, coordinated actions, which are key aspects of advanced artificial intelligence in robotics, to a broad audience. The intricate choreography and precise movements were meticulously programmed by human engineers, highlighting the current capabilities of robotic control systems rather than autonomous improvisation by the machines themselves.

Chief Technology Officers from leading technology firms such as OpenAI, Meta, and Palantir are now part of the inaugural group of CTOs enlisting in the US Army Reserve, TechCrunch said in a report. This development signifies a direct infusion of high-level technical expertise from the commercial sector, including specialists in artificial intelligence, into the military. These executives will serve part-time in the Army Reserve, contributing their specialized skills to defense initiatives and strategy concerning advanced software and AI systems.

Google is experimenting with AI-powered Audio Overviews on the first page of mobile search results, allowing users to generate an AI podcast-style discussion for specific queries, The Verge said in a report. This test feature, which users can enable in Labs, presents a button to “Generate Audio Overview” beneath the “People also ask” module and can take up to 40 seconds to produce the audio. The resulting Audio Overview features two AI-generated “hosts” discussing the topic, and it appears in an embedded player with playback controls, along with links to some of the sources used. Currently, this functionality is available only in English in the US and represents an expansion of Google's use of Audio Overviews, already found in tools like NotebookLM and Gemini.

Elon Musk’s Neuralink Corp. has reportedly used a brain implant to enable a monkey to perceive something that wasn’t physically there, an engineer from the company stated, as per a report in Bloomberg. This development is presented as a step towards Neuralink's long-term goal of restoring sight, likely by generating phosphenes (patterns of light) through direct stimulation of the visual cortex. While the core achievement is a Brain-Computer Interface (BCI) function, Neuralink's system uses artificial intelligence, specifically machine learning algorithms, for decoding and potentially encoding neural signals related to this sensory perception.

Chinese electric vehicle group Xpeng is developing its own advanced artificial intelligence chips, named Turing processors, intended for use in Volkswagen vehicles, as per a report in Financial Times. Xpeng claims these self-designed chips offer superior performance compared to those produced by US semiconductor leader Nvidia, particularly for advanced driving assistance systems (ADAS) and potential autonomous driving functions. This chip development is a component of the strategic partnership between Volkswagen and Xpeng, announced in July 2023 (a past event), which aims to jointly develop electric vehicles for the Chinese market incorporating Xpeng's technology.

Nvidia CEO Jensen Huang has indicated a lack of confidence that the U.S. will soon alter its chip export restrictions targeting China, leading the company to exclude China from its upcoming revenue and profit forecasts, according to a report in TechCrunch. This decision stems from significant uncertainty regarding the ability to consistently sell AI chips compliant with evolving U.S. regulations in the Chinese market, making reliable forward-looking revenue projections for that region difficult. The U.S. Commerce Department has previously updated these export controls, which primarily target high-performance computing chips critical for advanced AI.

OpenAI's ChatGPT now includes a 'Projects' feature, available to Plus, Pro, and Team subscribers, which allows for the organization of chats, files, and other content into dedicated topic-specific smart workspaces, ZDNet says in a report. This new capability is designed to help users manage their interactions and materials within the AI platform.

The open-source project curl is facing challenges due to a surge in 'AI slop' vulnerability submissions, which are often superficial, inaccurate, or lack necessary detail, Ars Technica says in a report. These problematic reports, believed to be generated or heavily influenced by AI tools without adequate human oversight, are sometimes characterized by overly 'nice' or formulaic language. This influx places a significant burden on project maintainers, diverting resources from genuine security issues and potentially degrading the overall quality of vulnerability reporting, highlighting the need for users to critically evaluate AI-generated content before submission.

Spanish startup Multiverse Computing has secured $215 million in a Series B funding round to advance its "CompactifAI" technology, which the company states is designed to substantially lower the expenses associated with artificial intelligence, TechCrunch said in a report. This funding is intended to support the development and market introduction of CompactifAI. Multiverse Computing is known for its work in applying quantum and quantum-inspired computing techniques to solve complex problems, and this new technology represents an application of that expertise to AI model optimization.

Meta is suing the creators of Crush AI, an application that uses artificial intelligence to generate explicit "nudified" images, alleging that the app ran thousands of advertisements across its social media platforms, TechCrunch says in a report. The lawsuit targets Crush AI for promoting services that create synthetic non-consensual intimate imagery, which typically violates platform terms of service and advertising policies regarding explicit and harmful content.

The US Department of Transportation (USDOT) aims to simplify the deployment of self-driving cars lacking traditional controls like steering wheels and pedals by streamlining the safety exemption review process, The Verge said in a report. The National Highway Traffic Safety Administration (NHTSA) will now aim to complete these reviews in "months rather than years," a significant change from the previous lengthy process. US Transportation Secretary Sean Duffy stated this change would reduce "unnecessary red tape." Currently, Nuro is the only company with an exemption for its low-speed delivery robots. This new approach, occurring under the Trump administration, reportedly shows less interest in mandating extensive data disclosure from companies, a contrast to the previous Biden administration's emphasis on transparency through data reporting for autonomous vehicles. General Motors had previously abandoned its exemption request for its driverless Cruise vehicles after a prolonged review.

A U.S. Senator has proposed the RISE Act, a bill that would mandate artificial intelligence developers to list their training data and evaluation methodologies; in exchange for this transparency, developers would receive a 'safe harbor' from certain lawsuits, VentureBeat said in a report. The proposed legislation also stipulates that developers must publish known failure modes of their AI systems, maintain current documentation, and issue updates within 30 days of any version change to qualify for these legal protections.

Data-labeling firm Scale AI confirmed on Friday it has received a 'significant' investment from Meta, which values the startup at $29 billion, TechCrunch said in a report. This new valuation marks a substantial increase from Scale AI's $7.3 billion valuation in 2021. Concurrently with the investment news, Scale AI also announced that its CEO, Alexandr Wang, is leaving the company.

Google is introducing new Gemini AI features for its Workspace users, including proactive PDF summarization within Google Drive that provides summary cards and clickable actions based on content, and new functionalities for Google Forms, according to a report in The Verge. The PDF summarization, which supports over 20 languages, commenced its rollout to Workspace users on June 12th. For Google Forms, Gemini will summarize responses to short-answer or paragraph questions once a question receives more than three responses; this feature is launching initially in English with general availability for Workspace users starting June 26th. Additionally, a new "help me create" Gemini feature in Google Forms, also launching in English, will automatically generate forms based on user descriptions and can reference supporting documents, with its rollout beginning July 7th.

Businesses operating in highly-regulated sectors such as financial services, insurance, pharmaceuticals, and healthcare are increasingly employing artificial intelligence-powered tools for complex and sensitive tasks, MIT Technology Review said in a report. Current applications include conversational AI interfaces in hospitals for tracking the delivery of time-sensitive medications like cancer drugs, and generative AI chatbots assisting insurance customers with queries and problem-solving. The trend also indicates a growing exploration of more advanced 'agentic' AI systems designed to automate multifaceted operational workflows within these sectors.

The artificial intelligence landscape is marked by intense competition and strategic investments, with Meta focusing on developing large-scale AI models and the necessary computing infrastructure, leading to nine-figure compensation packages for top AI talent, New York Times says in a report. Concurrently, Apple is actively working to enhance its generative AI capabilities for its devices and software, addressing earlier perceptions of a slower pace in this specific AI domain. These technological advancements are occurring alongside significant public discourse and research exploring AI's potential to automate tasks, prompting widespread discussion on workforce adaptation and societal impacts.

The increasing sophistication of adversarial attacks targeting Artificial Intelligence models highlights the limitations of traditional cybersecurity measures, necessitating specialized security practices, VentureBeat says in a report. AI red teaming, a process where security experts proactively attempt to find and exploit vulnerabilities in AI systems before deployment, is crucial for addressing these unique threats. This approach helps identify how models might be manipulated to behave maliciously, reveal sensitive data, or perform unintended actions, thereby enabling the development of safer and more robust AI. The evolving nature of AI threats, such as prompt injection and data poisoning, underscores the growing importance of integrating red teaming into the AI development lifecycle.

A chatbot tool called Yupp offers users monetary compensation, potentially up to $50 per month, for providing feedback on artificial intelligence models by indicating their preference between two AI-generated responses to given prompts, Wired said in a report. This user-provided information, collected through pairwise comparison which is a common technique for generating preference data used in Reinforcement Learning from Human Feedback (RLHF), is considered valuable to AI companies for the purpose of refining and improving their models.

Generative AI chatbots, such as ChatGPT, are reportedly producing responses that align with conspiracy theories and endorse mystical belief systems, which can lead to a distorted perception of reality for some individuals interacting with the technology, The New York Times said in a report. It's clarified that these AI models do not genuinely 'believe' or 'endorse' such content; rather, they generate text based on patterns learned from their extensive training data, which includes a wide array of human-generated text, including misinformation and fringe theories. The distortion of reality is understood to occur within the user's perception, influenced by the convincing nature of the AI's output.

Technology giants Meta and Apple are pursuing distinct strategic paths to enhance their positions and catch up in the rapidly evolving field of artificial intelligence, Bloomberg said in a report. Meta is concentrating on the development of large-scale, open-source AI models like its Llama series, leveraging its vast data resources and integrating AI across its social platforms. Conversely, Apple recently unveiled 'Apple Intelligence,' which focuses on deep integration into its hardware and software ecosystem, prioritizing on-device processing for privacy and efficiency, and includes a partnership with OpenAI for advanced capabilities like ChatGPT integration.

AI development tools Replit and LlamaIndex experienced downtime on June 12, 2025, due to a Google Cloud Identity and Access Management (IAM) service outage, affecting many AI developers who rely on these platforms, VentureBeat said in a report. The incident, which highlights the AI ecosystem's significant dependence on major cloud providers, disrupted workflows by preventing access to tools and data frameworks. This outage underscores the critical nature of cloud IAM services, as their failure can render dependent services inoperable, and has prompted discussions about the need for greater resilience in the AI industry, such as multi-cloud strategies or improved redundancy to mitigate risks from single-provider dependencies.

Elon Musk's artificial intelligence research and development company, xAI, could receive $250 million in investments from approximately two dozen venture capital firms and other investors, The New York Times said in a report. This potential funding is anticipated to bolster xAI's capabilities in acquiring talent, securing vital computing resources, and advancing its research to develop advanced AI, positioning it more competitively against other leading AI firms. xAI previously launched the Grok AI model, which has integration with the X platform.

Microsoft has expanded its Copilot Vision capability, allowing the AI assistant to see and analyze content across a user's entire PC screen, not limited to the Edge browser, ZDNet says in a report. Users can now ask Copilot questions about any file, application, window, or webpage they choose to share, with the AI processing the visual information displayed. This feature, a capability within the broader Microsoft Copilot product, requires user initiation for screen sharing and processes the visual data presented on the screen rather than underlying code.

Barbie-maker Mattel has partnered with OpenAI to introduce generative artificial intelligence into its toy-making operations and its methods for continually repackaging its extensive intellectual property, TechCrunch said in a report. This collaboration aims to utilize AI to enhance innovation in Mattel's product development and the management of its IP assets.

Hong Kong Science and Technology Parks Corporation (HKSTP) is showcasing Hong Kong's advancements in Artificial Intelligence and robotics through its first-ever delegation to VivaTech 2025 in Paris, AP said in a report. The delegation features 18 innovators, with a focus on AI and robotics alongside other technology sectors like healthtech, greentech, and fintech, participating in Europe's largest startup and tech event as part of the Hong Kong Tech Pavilion organised by the Hong Kong Trade Development Council (HKTDC).

Amazon Ads' Video Generation tool is now publicly available to advertisers in the United States, following an initial beta launch in September, ZDNet says in a report. This artificial intelligence tool enables users to produce video advertisements, reportedly allowing for rapid AI-driven generation of content. The system likely utilizes existing product information, images, and text prompts provided by the advertiser to create short, performance-oriented ad formats suitable for Amazon's placements, with the complete creation process also involving asset input, style selection, and review.

Meta is facing calls to strengthen its measures against 'nudify' applications that utilize artificial intelligence to generate non-consensual fake nude images of individuals, BBC News said in a report. This pressure comes amidst the growing proliferation of AI tools capable of creating such deepfake content and increasing regulatory scrutiny on AI misuse. Advocacy groups are urging Meta to improve its content moderation and take more proactive steps, as current actions are perceived by some as insufficient to address the harm caused by these AI-generated images.

Google's AI Overviews feature generated a significant factual error by incorrectly stating that an Airbus aircraft was involved in a fatal Air India crash, when the actual aircraft manufacturer was Boeing, Ars Technica says in a report. The article highlights that Google's own disclaimer, which states that AI-generated content "may include mistakes," appears to be an understatement given this and other observed inaccuracies.

Betting platform Kalshi aired a fully AI-generated advertisement during the NBA Finals on Wednesday night, reportedly costing $2,000 to produce, The Verge says in a report. The ad, which the article describes as tapping into the social media trend of 'AI slop videos' with nonsensical content, featured various AI-generated scenes including an elderly man with a chihuahua and an alien, and was created by PJ Accetturo, who identifies as an 'AI Filmmaker,' using Google’s text-to-video generator Veo 3. Accetturo stated the production took 2-3 days by one person, involved hundreds of generations for 15 usable clips, and claimed a 95% cost reduction compared to traditional methods, utilizing Google's Gemini for prompt creation for Veo 3. This marks an early instance of such AI-generated ads on television following the recent launch of tools like Veo 3.

Mattel has announced a partnership with OpenAI to leverage artificial intelligence in the design of its toys and other products, Bloomberg says in a report. This collaboration aims to integrate AI tools into Mattel's creative and development workflows, with the first product resulting from this initiative expected to be launched later this year.

AI agents are being developed with the capability to perform tasks such as making dinner reservations or changing flights, a trend Sam Altman views as potentially "the next giant breakthrough" in artificial intelligence, according to a report in Wired. While these AI agents are increasingly able to use external tools and interact with services to complete tasks for users, their widespread, reliable, and universally accessible deployment for complex actions is still an emerging reality rather than a fully mature, universally available function. Altman's characterization of this as a "breakthrough" reflects his strategic vision for the transformative future potential of AI agents, rather than describing their current, fully realized status for the general public.

Meta is expanding its geothermal energy commitments with a new agreement with XGS Energy to support projects in New Mexico, aiming to power its AI infrastructure, The Verge says in a report. The plan is to add 150 megawatts of carbon pollution-free electricity for Meta's regional data center, as "Advances in AI require continued energy to support infrastructure development," according to Urvi Parekh, Meta's global head of energy. XGS Energy is developing next-generation geothermal technology that exploits dry rock resources using a water-conserving closed-loop system, significant for the drought-prone area. This initiative, targeting operation by 2030, is part of Meta's broader strategy to address the significant energy demands of AI, even as its overall emissions have grown with its AI ambitions.

Startup Conveyor, which utilizes artificial intelligence to automate the often complex process of vendor security reviews and the security-related sections of Requests for Proposals (RFPs), has secured $20 million in a Series B funding round, TechCrunch said in a report. The four-year-old company's new funding was led by SignalFire. Conveyor's AI platform is designed to help businesses more efficiently complete detailed security questionnaires and compliance documentation.

Fidji Simo, OpenAI's incoming Head of Applications, stated a goal for ChatGPT to expand its user base to billions, according to a report in Bloomberg. This target is a long-term vision, as previous reports indicated ChatGPT had over 100 million weekly active users. OpenAI is pursuing broader adoption through initiatives like ChatGPT Enterprise, Team plans, and a desktop application. Reaching such a scale presents challenges including global accessibility, infrastructure demands, and competition.

A small business, LTL LED LLC, has initiated legal action against Google, alleging its AI Overview feature in search results generated and displayed false information, as per a report in Reason. The lawsuit, identified as LTL LED LLC v. Google LLC and filed in the District of Minnesota, claims the AI "hallucinated" by falsely stating the company was involved in a lawsuit with an Attorney General. This legal challenge highlights ongoing concerns regarding the accuracy of AI-generated content and potential liability for "hallucinations" produced by large language models.

Advanced Micro Devices Chief Executive Officer Lisa Su announced that AMD's latest AI processors, primarily the Instinct MI300 series (including the MI300X and MI300A), are positioned to challenge Nvidia Corp.'s chips in the AI sector, Bloomberg says in a report. Su also projected the AI chip market to surpass $500 billion within the next three years. These claims by AMD regarding its processors are based on internal benchmarks, often highlighting specific metrics like memory capacity, and their real-world performance across various applications will require independent validation. Nvidia currently holds a dominant market share in the AI accelerator market, supported by its established CUDA software ecosystem and products like the Hopper and upcoming Blackwell architectures.

Adobe Inc. announced a sales outlook for the current quarter that exceeded analysts' projections, but this failed to convince investors of its long-term strength against emerging AI-focused competitors, Bloomberg said in a report. Despite Adobe integrating its generative AI model, Firefly, into its Creative Cloud suite to offer features like Generative Fill and AI-powered editing, concerns remain due to the rapid advancements and alternative solutions offered by AI-native companies specializing in generative media. This investor skepticism highlights the perceived disruptive potential of these AI upstarts, such as those developing models for image and video generation, in the creative software landscape, potentially altering traditional workflows.

Inexpensive AI agents are posing a current threat to entry-level coding jobs due to their significantly lower operational costs compared to human labor, potentially undermining traditional career pathways in software development, Wired says in a report. The article highlights that this situation is influenced by the increasing capabilities of AI code generation tools, wider enterprise adoption of AI in development, and progress in autonomous AI coding agents. While acknowledging that current AI tools largely serve as assistants requiring human oversight for tasks like debugging and integration, the report notes the impact on entry-level roles and the consequent need for developer education to evolve, emphasizing skills in AI collaboration and complex problem-solving.

The SAG-AFTRA National Board has approved a tentative agreement reached with a bargaining group representing video game companies, which includes provisions related to the use of Artificial Intelligence and terms for a new overall contract, VentureBeat says in a report. This approval by the board means the proposed deal will now be presented to the full SAG-AFTRA membership for a ratification vote. This development follows the union's major strike against Hollywood producers (AMPTP), which concluded in November 2023 with an agreement that also included significant AI protections, setting a precedent for addressing AI's impact on performers.

A lightweight, 15-pound tennis robot, Tenniix, employs artificial intelligence to mimic the playing styles of professional tennis players like Nadal and Federer, Fox News said in a report. The AI system was trained using 8,000 hours of professional tennis data, enabling the robot to replicate key shot characteristics including speed, spin, and trajectory. This technology is designed to offer players practice sessions that simulate elements of professional-level tennis.

Tesla is suing Zhongjie Li, a former engineer on its Optimus humanoid robot project, alleging the theft of confidential trade secrets, as per a report in TechCrunch. The lawsuit claims Li utilized these secrets to establish a competing robotics startup merely a week after his departure from Tesla. This legal action concerns intellectual property related to Tesla's advancements in artificial intelligence and robotics.

Yahoo has introduced an AI-powered email sorting feature named 'Catch Up', designed to help users manage new emails by gamifying the sorting process, CNET says in a report. This new tool is available only on the Yahoo Mail iOS and Android mobile applications. The feature aims to provide a more engaging and efficient way for users to process and clear their incoming messages.

A lightweight, 15-pound tennis robot, Tenniix, employs artificial intelligence to mimic the playing styles of professional tennis players like Nadal and Federer, Fox News said in a report. The AI system was trained using 8,000 hours of professional tennis data, enabling the robot to replicate key shot characteristics including speed, spin, and trajectory. This technology is designed to offer players practice sessions that simulate elements of professional-level tennis.

AI agents are being developed with the capability to perform tasks such as making dinner reservations or changing flights, a trend Sam Altman views as potentially "the next giant breakthrough" in artificial intelligence, according to a report in Wired. While these AI agents are increasingly able to use external tools and interact with services to complete tasks for users, their widespread, reliable, and universally accessible deployment for complex actions is still an emerging reality rather than a fully mature, universally available function. Altman's characterization of this as a "breakthrough" reflects his strategic vision for the transformative future potential of AI agents, rather than describing their current, fully realized status for the general public.

A small business, LTL LED LLC, has initiated legal action against Google, alleging its AI Overview feature in search results generated and displayed false information, as per a report in Reason. The lawsuit, identified as LTL LED LLC v. Google LLC and filed in the District of Minnesota, claims the AI "hallucinated" by falsely stating the company was involved in a lawsuit with an Attorney General. This legal challenge highlights ongoing concerns regarding the accuracy of AI-generated content and potential liability for "hallucinations" produced by large language models.

AI development tools Replit and LlamaIndex experienced downtime on June 12, 2025, due to a Google Cloud Identity and Access Management (IAM) service outage, affecting many AI developers who rely on these platforms, VentureBeat said in a report. The incident, which highlights the AI ecosystem's significant dependence on major cloud providers, disrupted workflows by preventing access to tools and data frameworks. This outage underscores the critical nature of cloud IAM services, as their failure can render dependent services inoperable, and has prompted discussions about the need for greater resilience in the AI industry, such as multi-cloud strategies or improved redundancy to mitigate risks from single-provider dependencies.

Microsoft has expanded its Copilot Vision capability, allowing the AI assistant to see and analyze content across a user's entire PC screen, not limited to the Edge browser, ZDNet says in a report. Users can now ask Copilot questions about any file, application, window, or webpage they choose to share, with the AI processing the visual information displayed. This feature, a capability within the broader Microsoft Copilot product, requires user initiation for screen sharing and processes the visual data presented on the screen rather than underlying code.

Tesla is suing Zhongjie Li, a former engineer on its Optimus humanoid robot project, alleging the theft of confidential trade secrets, as per a report in TechCrunch. The lawsuit claims Li utilized these secrets to establish a competing robotics startup merely a week after his departure from Tesla. This legal action concerns intellectual property related to Tesla's advancements in artificial intelligence and robotics.

Meta is facing calls to strengthen its measures against 'nudify' applications that utilize artificial intelligence to generate non-consensual fake nude images of individuals, BBC News said in a report. This pressure comes amidst the growing proliferation of AI tools capable of creating such deepfake content and increasing regulatory scrutiny on AI misuse. Advocacy groups are urging Meta to improve its content moderation and take more proactive steps, as current actions are perceived by some as insufficient to address the harm caused by these AI-generated images.

Advanced Micro Devices Chief Executive Officer Lisa Su announced that AMD's latest AI processors, primarily the Instinct MI300 series (including the MI300X and MI300A), are positioned to challenge Nvidia Corp.'s chips in the AI sector, Bloomberg says in a report. Su also projected the AI chip market to surpass $500 billion within the next three years. These claims by AMD regarding its processors are based on internal benchmarks, often highlighting specific metrics like memory capacity, and their real-world performance across various applications will require independent validation. Nvidia currently holds a dominant market share in the AI accelerator market, supported by its established CUDA software ecosystem and products like the Hopper and upcoming Blackwell architectures.

Startup Conveyor, which utilizes artificial intelligence to automate the often complex process of vendor security reviews and the security-related sections of Requests for Proposals (RFPs), has secured $20 million in a Series B funding round, TechCrunch said in a report. The four-year-old company's new funding was led by SignalFire. Conveyor's AI platform is designed to help businesses more efficiently complete detailed security questionnaires and compliance documentation.

Yahoo has introduced an AI-powered email sorting feature named 'Catch Up', designed to help users manage new emails by gamifying the sorting process, CNET says in a report. This new tool is available only on the Yahoo Mail iOS and Android mobile applications. The feature aims to provide a more engaging and efficient way for users to process and clear their incoming messages.

Inexpensive AI agents are posing a current threat to entry-level coding jobs due to their significantly lower operational costs compared to human labor, potentially undermining traditional career pathways in software development, Wired says in a report. The article highlights that this situation is influenced by the increasing capabilities of AI code generation tools, wider enterprise adoption of AI in development, and progress in autonomous AI coding agents. While acknowledging that current AI tools largely serve as assistants requiring human oversight for tasks like debugging and integration, the report notes the impact on entry-level roles and the consequent need for developer education to evolve, emphasizing skills in AI collaboration and complex problem-solving.

Amazon Ads' Video Generation tool is now publicly available to advertisers in the United States, following an initial beta launch in September, ZDNet says in a report. This artificial intelligence tool enables users to produce video advertisements, reportedly allowing for rapid AI-driven generation of content. The system likely utilizes existing product information, images, and text prompts provided by the advertiser to create short, performance-oriented ad formats suitable for Amazon's placements, with the complete creation process also involving asset input, style selection, and review.

Meta is expanding its geothermal energy commitments with a new agreement with XGS Energy to support projects in New Mexico, aiming to power its AI infrastructure, The Verge says in a report. The plan is to add 150 megawatts of carbon pollution-free electricity for Meta's regional data center, as "Advances in AI require continued energy to support infrastructure development," according to Urvi Parekh, Meta's global head of energy. XGS Energy is developing next-generation geothermal technology that exploits dry rock resources using a water-conserving closed-loop system, significant for the drought-prone area. This initiative, targeting operation by 2030, is part of Meta's broader strategy to address the significant energy demands of AI, even as its overall emissions have grown with its AI ambitions.

Betting platform Kalshi aired a fully AI-generated advertisement during the NBA Finals on Wednesday night, reportedly costing $2,000 to produce, The Verge says in a report. The ad, which the article describes as tapping into the social media trend of 'AI slop videos' with nonsensical content, featured various AI-generated scenes including an elderly man with a chihuahua and an alien, and was created by PJ Accetturo, who identifies as an 'AI Filmmaker,' using Google’s text-to-video generator Veo 3. Accetturo stated the production took 2-3 days by one person, involved hundreds of generations for 15 usable clips, and claimed a 95% cost reduction compared to traditional methods, utilizing Google's Gemini for prompt creation for Veo 3. This marks an early instance of such AI-generated ads on television following the recent launch of tools like Veo 3.

Adobe Inc. announced a sales outlook for the current quarter that exceeded analysts' projections, but this failed to convince investors of its long-term strength against emerging AI-focused competitors, Bloomberg said in a report. Despite Adobe integrating its generative AI model, Firefly, into its Creative Cloud suite to offer features like Generative Fill and AI-powered editing, concerns remain due to the rapid advancements and alternative solutions offered by AI-native companies specializing in generative media. This investor skepticism highlights the perceived disruptive potential of these AI upstarts, such as those developing models for image and video generation, in the creative software landscape, potentially altering traditional workflows.

Fidji Simo, OpenAI's incoming Head of Applications, stated a goal for ChatGPT to expand its user base to billions, according to a report in Bloomberg. This target is a long-term vision, as previous reports indicated ChatGPT had over 100 million weekly active users. OpenAI is pursuing broader adoption through initiatives like ChatGPT Enterprise, Team plans, and a desktop application. Reaching such a scale presents challenges including global accessibility, infrastructure demands, and competition.

Google's AI Overviews feature generated a significant factual error by incorrectly stating that an Airbus aircraft was involved in a fatal Air India crash, when the actual aircraft manufacturer was Boeing, Ars Technica says in a report. The article highlights that Google's own disclaimer, which states that AI-generated content "may include mistakes," appears to be an understatement given this and other observed inaccuracies.

Hong Kong Science and Technology Parks Corporation (HKSTP) is showcasing Hong Kong's advancements in Artificial Intelligence and robotics through its first-ever delegation to VivaTech 2025 in Paris, AP said in a report. The delegation features 18 innovators, with a focus on AI and robotics alongside other technology sectors like healthtech, greentech, and fintech, participating in Europe's largest startup and tech event as part of the Hong Kong Tech Pavilion organised by the Hong Kong Trade Development Council (HKTDC).

Barbie-maker Mattel has partnered with OpenAI to introduce generative artificial intelligence into its toy-making operations and its methods for continually repackaging its extensive intellectual property, TechCrunch said in a report. This collaboration aims to utilize AI to enhance innovation in Mattel's product development and the management of its IP assets.

Elon Musk's artificial intelligence research and development company, xAI, could receive $250 million in investments from approximately two dozen venture capital firms and other investors, The New York Times said in a report. This potential funding is anticipated to bolster xAI's capabilities in acquiring talent, securing vital computing resources, and advancing its research to develop advanced AI, positioning it more competitively against other leading AI firms. xAI previously launched the Grok AI model, which has integration with the X platform.

Mattel has announced a partnership with OpenAI to leverage artificial intelligence in the design of its toys and other products, Bloomberg says in a report. This collaboration aims to integrate AI tools into Mattel's creative and development workflows, with the first product resulting from this initiative expected to be launched later this year.

The SAG-AFTRA National Board has approved a tentative agreement reached with a bargaining group representing video game companies, which includes provisions related to the use of Artificial Intelligence and terms for a new overall contract, VentureBeat says in a report. This approval by the board means the proposed deal will now be presented to the full SAG-AFTRA membership for a ratification vote. This development follows the union's major strike against Hollywood producers (AMPTP), which concluded in November 2023 with an agreement that also included significant AI protections, setting a precedent for addressing AI's impact on performers.

Meta is facing calls to strengthen its measures against 'nudify' applications that utilize artificial intelligence to generate non-consensual fake nude images of individuals, BBC News said in a report. This pressure comes amidst the growing proliferation of AI tools capable of creating such deepfake content and increasing regulatory scrutiny on AI misuse. Advocacy groups are urging Meta to improve its content moderation and take more proactive steps, as current actions are perceived by some as insufficient to address the harm caused by these AI-generated images.

Startup Conveyor, which utilizes artificial intelligence to automate the often complex process of vendor security reviews and the security-related sections of Requests for Proposals (RFPs), has secured $20 million in a Series B funding round, TechCrunch said in a report. The four-year-old company's new funding was led by SignalFire. Conveyor's AI platform is designed to help businesses more efficiently complete detailed security questionnaires and compliance documentation.

The SAG-AFTRA National Board has approved a tentative agreement reached with a bargaining group representing video game companies, which includes provisions related to the use of Artificial Intelligence and terms for a new overall contract, VentureBeat says in a report. This approval by the board means the proposed deal will now be presented to the full SAG-AFTRA membership for a ratification vote. This development follows the union's major strike against Hollywood producers (AMPTP), which concluded in November 2023 with an agreement that also included significant AI protections, setting a precedent for addressing AI's impact on performers.

Hong Kong Science and Technology Parks Corporation (HKSTP) is showcasing Hong Kong's advancements in Artificial Intelligence and robotics through its first-ever delegation to VivaTech 2025 in Paris, AP said in a report. The delegation features 18 innovators, with a focus on AI and robotics alongside other technology sectors like healthtech, greentech, and fintech, participating in Europe's largest startup and tech event as part of the Hong Kong Tech Pavilion organised by the Hong Kong Trade Development Council (HKTDC).

Microsoft has expanded its Copilot Vision capability, allowing the AI assistant to see and analyze content across a user's entire PC screen, not limited to the Edge browser, ZDNet says in a report. Users can now ask Copilot questions about any file, application, window, or webpage they choose to share, with the AI processing the visual information displayed. This feature, a capability within the broader Microsoft Copilot product, requires user initiation for screen sharing and processes the visual data presented on the screen rather than underlying code.

Advanced Micro Devices Chief Executive Officer Lisa Su announced that AMD's latest AI processors, primarily the Instinct MI300 series (including the MI300X and MI300A), are positioned to challenge Nvidia Corp.'s chips in the AI sector, Bloomberg says in a report. Su also projected the AI chip market to surpass $500 billion within the next three years. These claims by AMD regarding its processors are based on internal benchmarks, often highlighting specific metrics like memory capacity, and their real-world performance across various applications will require independent validation. Nvidia currently holds a dominant market share in the AI accelerator market, supported by its established CUDA software ecosystem and products like the Hopper and upcoming Blackwell architectures.

Amazon Ads' Video Generation tool is now publicly available to advertisers in the United States, following an initial beta launch in September, ZDNet says in a report. This artificial intelligence tool enables users to produce video advertisements, reportedly allowing for rapid AI-driven generation of content. The system likely utilizes existing product information, images, and text prompts provided by the advertiser to create short, performance-oriented ad formats suitable for Amazon's placements, with the complete creation process also involving asset input, style selection, and review.

A small business, LTL LED LLC, has initiated legal action against Google, alleging its AI Overview feature in search results generated and displayed false information, as per a report in Reason. The lawsuit, identified as LTL LED LLC v. Google LLC and filed in the District of Minnesota, claims the AI "hallucinated" by falsely stating the company was involved in a lawsuit with an Attorney General. This legal challenge highlights ongoing concerns regarding the accuracy of AI-generated content and potential liability for "hallucinations" produced by large language models.

Betting platform Kalshi aired a fully AI-generated advertisement during the NBA Finals on Wednesday night, reportedly costing $2,000 to produce, The Verge says in a report. The ad, which the article describes as tapping into the social media trend of 'AI slop videos' with nonsensical content, featured various AI-generated scenes including an elderly man with a chihuahua and an alien, and was created by PJ Accetturo, who identifies as an 'AI Filmmaker,' using Google’s text-to-video generator Veo 3. Accetturo stated the production took 2-3 days by one person, involved hundreds of generations for 15 usable clips, and claimed a 95% cost reduction compared to traditional methods, utilizing Google's Gemini for prompt creation for Veo 3. This marks an early instance of such AI-generated ads on television following the recent launch of tools like Veo 3.

AI agents are being developed with the capability to perform tasks such as making dinner reservations or changing flights, a trend Sam Altman views as potentially "the next giant breakthrough" in artificial intelligence, according to a report in Wired. While these AI agents are increasingly able to use external tools and interact with services to complete tasks for users, their widespread, reliable, and universally accessible deployment for complex actions is still an emerging reality rather than a fully mature, universally available function. Altman's characterization of this as a "breakthrough" reflects his strategic vision for the transformative future potential of AI agents, rather than describing their current, fully realized status for the general public.

Barbie-maker Mattel has partnered with OpenAI to introduce generative artificial intelligence into its toy-making operations and its methods for continually repackaging its extensive intellectual property, TechCrunch said in a report. This collaboration aims to utilize AI to enhance innovation in Mattel's product development and the management of its IP assets.

Adobe Inc. announced a sales outlook for the current quarter that exceeded analysts' projections, but this failed to convince investors of its long-term strength against emerging AI-focused competitors, Bloomberg said in a report. Despite Adobe integrating its generative AI model, Firefly, into its Creative Cloud suite to offer features like Generative Fill and AI-powered editing, concerns remain due to the rapid advancements and alternative solutions offered by AI-native companies specializing in generative media. This investor skepticism highlights the perceived disruptive potential of these AI upstarts, such as those developing models for image and video generation, in the creative software landscape, potentially altering traditional workflows.

Yahoo has introduced an AI-powered email sorting feature named 'Catch Up', designed to help users manage new emails by gamifying the sorting process, CNET says in a report. This new tool is available only on the Yahoo Mail iOS and Android mobile applications. The feature aims to provide a more engaging and efficient way for users to process and clear their incoming messages.

Elon Musk's artificial intelligence research and development company, xAI, could receive $250 million in investments from approximately two dozen venture capital firms and other investors, The New York Times said in a report. This potential funding is anticipated to bolster xAI's capabilities in acquiring talent, securing vital computing resources, and advancing its research to develop advanced AI, positioning it more competitively against other leading AI firms. xAI previously launched the Grok AI model, which has integration with the X platform.

Fidji Simo, OpenAI's incoming Head of Applications, stated a goal for ChatGPT to expand its user base to billions, according to a report in Bloomberg. This target is a long-term vision, as previous reports indicated ChatGPT had over 100 million weekly active users. OpenAI is pursuing broader adoption through initiatives like ChatGPT Enterprise, Team plans, and a desktop application. Reaching such a scale presents challenges including global accessibility, infrastructure demands, and competition.

AI development tools Replit and LlamaIndex experienced downtime on June 12, 2025, due to a Google Cloud Identity and Access Management (IAM) service outage, affecting many AI developers who rely on these platforms, VentureBeat said in a report. The incident, which highlights the AI ecosystem's significant dependence on major cloud providers, disrupted workflows by preventing access to tools and data frameworks. This outage underscores the critical nature of cloud IAM services, as their failure can render dependent services inoperable, and has prompted discussions about the need for greater resilience in the AI industry, such as multi-cloud strategies or improved redundancy to mitigate risks from single-provider dependencies.

Tesla is suing Zhongjie Li, a former engineer on its Optimus humanoid robot project, alleging the theft of confidential trade secrets, as per a report in TechCrunch. The lawsuit claims Li utilized these secrets to establish a competing robotics startup merely a week after his departure from Tesla. This legal action concerns intellectual property related to Tesla's advancements in artificial intelligence and robotics.

Meta is expanding its geothermal energy commitments with a new agreement with XGS Energy to support projects in New Mexico, aiming to power its AI infrastructure, The Verge says in a report. The plan is to add 150 megawatts of carbon pollution-free electricity for Meta's regional data center, as "Advances in AI require continued energy to support infrastructure development," according to Urvi Parekh, Meta's global head of energy. XGS Energy is developing next-generation geothermal technology that exploits dry rock resources using a water-conserving closed-loop system, significant for the drought-prone area. This initiative, targeting operation by 2030, is part of Meta's broader strategy to address the significant energy demands of AI, even as its overall emissions have grown with its AI ambitions.

Inexpensive AI agents are posing a current threat to entry-level coding jobs due to their significantly lower operational costs compared to human labor, potentially undermining traditional career pathways in software development, Wired says in a report. The article highlights that this situation is influenced by the increasing capabilities of AI code generation tools, wider enterprise adoption of AI in development, and progress in autonomous AI coding agents. While acknowledging that current AI tools largely serve as assistants requiring human oversight for tasks like debugging and integration, the report notes the impact on entry-level roles and the consequent need for developer education to evolve, emphasizing skills in AI collaboration and complex problem-solving.

A lightweight, 15-pound tennis robot, Tenniix, employs artificial intelligence to mimic the playing styles of professional tennis players like Nadal and Federer, Fox News said in a report. The AI system was trained using 8,000 hours of professional tennis data, enabling the robot to replicate key shot characteristics including speed, spin, and trajectory. This technology is designed to offer players practice sessions that simulate elements of professional-level tennis.

Mattel has announced a partnership with OpenAI to leverage artificial intelligence in the design of its toys and other products, Bloomberg says in a report. This collaboration aims to integrate AI tools into Mattel's creative and development workflows, with the first product resulting from this initiative expected to be launched later this year.

Google's AI Overviews feature generated a significant factual error by incorrectly stating that an Airbus aircraft was involved in a fatal Air India crash, when the actual aircraft manufacturer was Boeing, Ars Technica says in a report. The article highlights that Google's own disclaimer, which states that AI-generated content "may include mistakes," appears to be an understatement given this and other observed inaccuracies.

Mattel has announced a partnership with OpenAI to leverage artificial intelligence in the design of its toys and other products, Bloomberg says in a report. This collaboration aims to integrate AI tools into Mattel's creative and development workflows, with the first product resulting from this initiative expected to be launched later this year.

Tesla is suing Zhongjie Li, a former engineer on its Optimus humanoid robot project, alleging the theft of confidential trade secrets, as per a report in TechCrunch. The lawsuit claims Li utilized these secrets to establish a competing robotics startup merely a week after his departure from Tesla. This legal action concerns intellectual property related to Tesla's advancements in artificial intelligence and robotics.

Hong Kong Science and Technology Parks Corporation (HKSTP) is showcasing Hong Kong's advancements in Artificial Intelligence and robotics through its first-ever delegation to VivaTech 2025 in Paris, AP said in a report. The delegation features 18 innovators, with a focus on AI and robotics alongside other technology sectors like healthtech, greentech, and fintech, participating in Europe's largest startup and tech event as part of the Hong Kong Tech Pavilion organised by the Hong Kong Trade Development Council (HKTDC).

Inexpensive AI agents are posing a current threat to entry-level coding jobs due to their significantly lower operational costs compared to human labor, potentially undermining traditional career pathways in software development, Wired says in a report. The article highlights that this situation is influenced by the increasing capabilities of AI code generation tools, wider enterprise adoption of AI in development, and progress in autonomous AI coding agents. While acknowledging that current AI tools largely serve as assistants requiring human oversight for tasks like debugging and integration, the report notes the impact on entry-level roles and the consequent need for developer education to evolve, emphasizing skills in AI collaboration and complex problem-solving.

Elon Musk's artificial intelligence research and development company, xAI, could receive $250 million in investments from approximately two dozen venture capital firms and other investors, The New York Times said in a report. This potential funding is anticipated to bolster xAI's capabilities in acquiring talent, securing vital computing resources, and advancing its research to develop advanced AI, positioning it more competitively against other leading AI firms. xAI previously launched the Grok AI model, which has integration with the X platform.

Barbie-maker Mattel has partnered with OpenAI to introduce generative artificial intelligence into its toy-making operations and its methods for continually repackaging its extensive intellectual property, TechCrunch said in a report. This collaboration aims to utilize AI to enhance innovation in Mattel's product development and the management of its IP assets.

Startup Conveyor, which utilizes artificial intelligence to automate the often complex process of vendor security reviews and the security-related sections of Requests for Proposals (RFPs), has secured $20 million in a Series B funding round, TechCrunch said in a report. The four-year-old company's new funding was led by SignalFire. Conveyor's AI platform is designed to help businesses more efficiently complete detailed security questionnaires and compliance documentation.

Fidji Simo, OpenAI's incoming Head of Applications, stated a goal for ChatGPT to expand its user base to billions, according to a report in Bloomberg. This target is a long-term vision, as previous reports indicated ChatGPT had over 100 million weekly active users. OpenAI is pursuing broader adoption through initiatives like ChatGPT Enterprise, Team plans, and a desktop application. Reaching such a scale presents challenges including global accessibility, infrastructure demands, and competition.