Latest AI News

Anthropic's law firm Latham & Watkins has admitted that its attorney used Claude AI to generate a citation in a music publishers' copyright lawsuit, resulting in a hallucinated reference with fake authors and title. The incident occurred when attorney Ivana Dukanovic used Anthropic's own AI to format a citation for a legitimate academic article, but failed to catch the errors during review. U.S. Magistrate Judge Susan van Keulen called the situation 'a very serious and grave issue,' highlighting growing concerns about AI reliability in legal contexts.

Cognichip has emerged from stealth mode with $33 million in seed funding to revolutionize semiconductor development through its Artificial Chip Intelligence (ACI®) platform. The San Francisco-based startup, founded by semiconductor veteran Faraj Aalaei, aims to reduce chip design time by 50% and cut development costs by 75% using a physics-informed AI foundation model. This technology could democratize access to chip design, allowing smaller companies to compete in a field traditionally dominated by large corporations.

Google has rolled out significant AI-powered accessibility enhancements to Android and Chrome, with TalkBack's Gemini integration being the standout feature. The update allows visually impaired users to not only receive AI-generated image descriptions but also ask follow-up questions about images and screen content. This advancement, available on Android 15 devices in select English-speaking countries, represents a major step forward in creating more inclusive technology.

Chinese humanoid robot companies like MagicLab and AgiBot are integrating advanced AI models to transform manufacturing capabilities. MagicLab CEO Wu Changzheng has successfully deployed robot prototypes for quality inspection, material handling, and assembly tasks by leveraging AI models from DeepSeek, Alibaba's Qwen, and ByteDance's Doubao. With government backing of over $20 billion and a national plan to mass-produce humanoid robots by 2025, China aims to address its projected labor shortage of nearly 30 million manufacturing workers.

FDA Commissioner Martin Makary announced an aggressive timeline to implement artificial intelligence across all FDA centers by June 30, 2025, following a successful generative AI pilot program. The initiative aims to dramatically reduce time spent on repetitive tasks, with officials reporting that processes previously taking days can now be completed in minutes. This historic move represents a significant shift from theoretical discussions to concrete implementation of AI in government regulatory processes.

UAE tech firm G42 has partnered with Italian AI unicorn iGenius to develop Colosseum, a major AI supercomputer in southern Italy. The $1 billion project, to be developed over five years using Nvidia's advanced Blackwell technology, aims to create Europe's largest AI computing infrastructure. This collaboration is part of the broader $40 billion UAE investment framework in Italy announced in February 2025 by Italian Prime Minister Giorgia Meloni.

President Donald Trump announced a landmark agreement allowing the UAE to import 500,000 advanced Nvidia AI chips annually starting in 2025, marking a major shift in US export policy. The deal centers around a massive 10-square-mile AI campus in Abu Dhabi with 5 gigawatts of power capacity, which Rand Corporation analyst Lennart Heim describes as 'bigger than all other major AI infrastructure announcements' seen so far. This agreement positions the UAE to become a significant global AI hub while addressing previous US concerns about technology transfers to China.

Wall Street rallied for a fifth consecutive day on Friday as markets continued to respond positively to the US-China tariff truce announced earlier this week. The agreement, which slashes reciprocal tariffs from 145% to 30% for the US and from 125% to 10% for China over a 90-day period, has particular significance for the AI industry by potentially easing tensions in the global semiconductor supply chain. Despite this positive development, economic survey data showed deteriorating consumer sentiment, suggesting underlying economic concerns remain.

Mike Novogratz's Galaxy Digital made its long-awaited Nasdaq debut on May 16, 2025, trading under ticker GLXY after a four-year regulatory battle. This milestone coincides with Coinbase's upcoming addition to the S&P 500 on May 19 and eToro's successful Nasdaq IPO earlier in the week. These developments signal growing mainstream acceptance of cryptocurrency and blockchain technologies, which are increasingly intersecting with AI applications.

Mike Novogratz's Galaxy Digital made its long-awaited Nasdaq debut on May 16, 2025, trading under the ticker GLXY at $23.50 per share after a four-year regulatory battle. The listing coincided with a landmark week for cryptocurrency in public markets, with Coinbase set to join the S&P 500 on May 19 and retail brokerage eToro also going public. Galaxy Digital positions itself at the intersection of cryptocurrency and artificial intelligence, with Novogratz describing the company as both 'a data center company and a crypto company.'

Elon Musk's xAI has confirmed that an unauthorized modification to its Grok chatbot on May 14, 2025, caused it to generate unprompted responses about 'white genocide' in South Africa. The company attributed the incident to a rogue employee who circumvented existing code review processes. In response, xAI announced several new security measures, including publishing Grok's system prompts on GitHub and establishing 24/7 monitoring of the chatbot's responses.

A bipartisan coalition of 40 state attorneys general formally opposed a Republican proposal to ban state-level AI regulation for 10 years on May 16, 2025. The controversial measure, included in President Trump's tax bill, would preempt dozens of existing state AI laws designed to protect consumers from potential harms. State officials argue the federal moratorium would create a dangerous regulatory void while failing to establish alternative protections.

The Federal Communications Commission has approved Verizon's $20 billion acquisition of Frontier Communications, marking a significant expansion of Verizon's fiber network infrastructure. The deal, which follows Verizon's agreement to revise its diversity policies, will strengthen the telecom giant's intelligent edge network capabilities for AI and IoT innovations. With this acquisition, Verizon aims to accelerate its position in the rapidly growing AI edge computing market, projected to reach $270 billion by 2032.

Epic Games announced on May 16, 2025 that Apple has blocked Fortnite from returning to the US App Store and removed it from the EU's Epic Games Store for iOS. This latest development follows Epic's recent court victory that forced Apple to allow external payment options without commissions. Apple claims it merely requested Epic Sweden resubmit the app without including the US storefront to avoid impacting availability in other regions.

OpenAI is set to become a primary anchor tenant in a sprawling 5-gigawatt data center campus in Abu Dhabi, potentially creating one of the world's largest AI infrastructure projects. The facility, spanning 10 square miles and requiring power equivalent to five nuclear reactors, is part of a broader US-UAE technology partnership announced during President Trump's Middle East tour. This development coincides with Trump's approval of advanced AI chip exports to the UAE, signaling a major shift in US technology transfer policy.

OpenAI is set to become a primary anchor tenant in a massive 5-gigawatt data center campus planned for Abu Dhabi, following a landmark agreement between the UAE and United States. The sprawling facility, to be built by Emirati firm G42, will span 10 square miles and potentially become one of the world's largest AI infrastructure projects. This development marks a significant expansion of OpenAI's global AI ambitions and highlights the Middle East's growing importance in the technology landscape.

Elon Musk's xAI has implemented new transparency measures after its Grok chatbot delivered unsolicited comments about 'white genocide' in South Africa due to an unauthorized system prompt modification. The company will now publish all system prompts on GitHub, allowing public review of changes, and establish additional safeguards to prevent similar incidents. This marks the second time xAI has acknowledged unauthorized changes to Grok's programming in 2025.

The European Investment Bank has launched an ambitious new project called 'Tech EU' to strengthen European capabilities in artificial intelligence and semiconductors, aiming to raise €70 billion ($78 billion) by 2027. This initiative, announced by EIB President Nadia Calviño, is part of the EU's broader strategy to enhance its technological competitiveness against China and the United States. The long-term goal is to attract a total of €250 billion in investments, with additional focus on health technologies and critical commodities.

U.S. tech giants Nvidia, Cisco, Oracle, and OpenAI are collaborating on the ambitious 'UAE Stargate' AI data center in Abu Dhabi, which will be built by Emirati firm G42. The 10-square-mile facility, announced during President Trump's Middle East visit, will eventually reach a staggering 5-gigawatt capacity. Nvidia will supply its latest Blackwell GB300 systems, representing a major expansion of global AI infrastructure and strengthening U.S.-UAE technology partnerships.

MIT researchers have discovered that vision-language models used in medical imaging cannot comprehend negation words like 'no' and 'not', potentially leading to dangerous misdiagnoses. When tested on negation tasks, these AI systems performed no better than random guessing, raising serious concerns about their deployment in healthcare settings. Researchers have developed a new benchmark called NegBench and proposed solutions that could improve negation understanding by up to 28%.