hello ai today15 headlines · 224 sources · published May 24 6:00 PM – May 25 6:00 AM
- 01
Pope Leo XIV Calls for AI to Be 'Disarmed' in Landmark Papal Encyclical
Pope Leo XIV has issued his first papal encyclical, 'Magnifica humanitas,' calling for artificial intelligence to be "disarmed" and warning that the technology risks concentrating power in the hands of a few and threatening humanity's future. In the sweeping 42,300-word document, the Pontiff compared AI's dangers to the biblical Tower of Babel, condemned what he called "the idolatry of profit," and urged that technology be made "human friendly." He warned of AI's harms in education, child development, and privacy. Notably, Anthropic's Chris Olah appeared alongside Pope Leo XIV at the Vatican launch of Magnifica humanitas, acknowledging that AI cannot be safely steered by tech labs alone. The encyclical calls for breaking up monopolistic control of AI and frames the issue as a moral and civilizational challenge, adding significant religious and cultural weight to the global AI governance debate.
- 02
SoftBank Shares Hit Record High on OpenAI IPO Hopes
SoftBank Group shares surged to a record high, driven by rising investor optimism around a potential OpenAI IPO and the broader AI investment boom. The Japanese conglomerate, which has made massive bets on AI through its Vision Fund and its recently announced $500 billion Stargate project with OpenAI, stands to reap enormous returns if OpenAI proceeds with a public offering. Separately, SoftBank is also tapping retail markets again with a ¥260 billion ($1.6 billion) yen bond issuance, signaling continued appetite for capital to fund its AI ambitions. SoftBank's record stock price underscores how closely its fortunes are now tied to the AI sector's trajectory — making it a bellwether for broader sentiment around AI valuations and the expected wave of AI-related IPOs in 2026.
- 03
AI Boom Fuels Record Run for Global Momentum Stocks; Big Tech Dominates Bond Markets
The AI investment boom is delivering the best return for global momentum stocks on record, according to Bloomberg, as investors pile into AI-linked equities and the halo effect spreads across the technology sector. Separately, the Financial Times reports that US tech giants are also dominating bond markets, with major companies like Anthropic and others racing to tap credit markets to fund the construction of AI data centers. The dual dynamic — soaring equity valuations and heavy bond issuance — reflects the unprecedented scale of capital flowing into AI infrastructure. Analysts warn that the momentum trade is increasingly concentrated, raising the risk of a sharp reversal if AI revenue growth disappoints. Meanwhile, UK venture funding doubled to $10.5 billion in the first four months of 2026, largely on the back of a few giant AI-related rounds, further illustrating how AI is reshaping global capital allocation.
- 04
AI Safety Guardrails on Meta and Google Models Stripped in Minutes, FT Reports
The Financial Times reports that software tools specifically designed to remove safety protections from major AI models — including those from Meta and Google — can bypass guardrails within minutes, enabling systems to provide harmful, dangerous, or illegal responses. The findings raise serious questions about the robustness of AI safety measures deployed by leading labs and the effectiveness of voluntary safety commitments made to regulators. As AI models proliferate across consumer and enterprise settings, the ease with which safety layers can be stripped away signals a fundamental gap between promised and actual protections. The report comes at a moment of heightened scrutiny of AI safety practices, with the Pope's encyclical and growing regulatory pressure in both the US and EU adding urgency to calls for binding, independently verified safety standards rather than self-policed guidelines.
- 05
Anthropic's Mythos AI Finds Over 10,000 Critical Security Vulnerabilities in Major Software
Anthropic has revealed that its Mythos AI security system, during its preview period, discovered over 10,000 critical- and high-severity vulnerabilities in some of the world's most systemically important software. Partners reported finding hundreds of critical vulnerabilities each within just one month of using the system. Separately, the European Central Bank has convened major European banks for an emergency meeting to address the cybersecurity weaknesses that AI models like Mythos keep uncovering, underscoring the systemic implications of AI-driven vulnerability discovery. The developments highlight both the transformative potential of AI in offensive and defensive cybersecurity and the urgent need for financial institutions to patch longstanding weaknesses before malicious actors can exploit them. The Wired report also notes a broader AI-driven bug-hunting arms race, with attackers also ramping up use of AI for exploit development.
- 06
California State University System Embraces AI Amid Faculty and Student Pushback
The California State University (CSU) system, one of the largest university networks in the United States, is moving to broadly embrace AI tools across its campuses, offering an early and high-profile look at what happens when a major public institution institutionalizes AI use. NPR reports that while administrators see AI as a way to improve learning outcomes and operational efficiency, faculty and students are far from unanimous in their support. Key concerns include academic integrity, the erosion of critical thinking skills, equity of access to AI tools, and the displacement of human instruction. CSU's experience is being closely watched as a potential template — or cautionary tale — for higher education institutions worldwide grappling with how to integrate AI responsibly. The tensions highlight a broader societal debate about who gets to decide how AI is adopted in public institutions.
- 07
Microsoft Engineers' Quiet Retreat from Claude Code Exposes Real Costs of Enterprise AI
A detailed report from The Next Web reveals that in December, Microsoft instructed thousands of its engineers and product managers to use Claude Code — Anthropic's AI coding assistant — as a primary development tool. But months later, Microsoft has quietly walked back the initiative, with engineers reporting friction, inconsistent results, and the hidden costs of integrating AI coding tools into complex enterprise workflows. The episode illuminates a growing gap between the marketing promise of AI-augmented software development and the on-the-ground reality faced by large engineering organizations. George Hotz, the programmer and entrepreneur, separately warned that AI coding agents will become "one of the most costly mistakes" in software development, arguing that agent-generated code is difficult to debug, maintain, and trust at scale. Together, the reports raise important questions about the true productivity gains from AI coding tools in enterprise settings.
- 08
Samsung Q2 Profit Expected to Soar on AI Memory Demand; Galaxy Z Fold 8 Ultra Leaks
Samsung Electronics is expected to post a significant surge in Q2 profit, driven by booming demand for AI memory chips — including HBM (High Bandwidth Memory) used in AI accelerators — as hyperscalers and AI companies continue to invest heavily in infrastructure. The AI memory cycle appears to be inflecting upward again after a period of uncertainty over HBM supply and demand dynamics. Separately, leaks have emerged for what is being called the 'Galaxy Z Fold 8 Ultra,' a new foldable variant that appears designed primarily as a competitive response to rivals rather than a mainstream product, suggesting Samsung is expanding its foldable lineup to hold ground against Chinese foldable makers. Together, the stories illustrate Samsung's dual strategy of capitalizing on the AI infrastructure boom in its semiconductor division while fighting intensifying competition in its mobile devices business.
- 09
Google DeepMind's AlphaProof Nexus Solves Decades-Old Math Problems for Hundreds of Dollars
Google DeepMind's AlphaProof Nexus has autonomously solved nine open Erdős problems — mathematical conjectures that have stumped researchers for decades — at a cost of just a few hundred dollars per problem. The breakthrough marks a significant leap in AI-powered mathematical reasoning, demonstrating that frontier AI systems can now crack longstanding open problems in combinatorics and number theory without human guidance. The Erdős problems have historically been considered among the most challenging unsolved questions in mathematics. AlphaProof Nexus's success builds on the earlier AlphaProof system, which made headlines by solving International Mathematical Olympiad problems. The development signals a new era for AI in scientific discovery, with potential implications for fields ranging from cryptography to theoretical physics, and underscores the accelerating pace at which AI is moving from tool to autonomous research collaborator.
- 10
California Governor Orders State to Prepare for AI-Driven Workforce Disruption
California Governor Gavin Newsom has issued an executive order directing state agencies and businesses to prepare for AI-driven workforce disruption. The order tasks state agencies with assessing the impact of AI on jobs across key sectors and developing strategies for workforce retraining and transition support. California, home to the world's largest concentration of AI companies, is positioning itself as a test case for proactive government response to automation-driven displacement. The move comes as concerns mount over Big Tech layoffs — which one analysis suggests are disproportionately affecting large tech companies — and broader anxiety about AI replacing knowledge workers. The executive order represents one of the most significant state-level policy interventions on AI and employment to date, potentially setting a template for other states and countries grappling with the same challenge.
- 11
AI Is Crashing Cannes Film Festival With Multi-Year Meta Partnership
Artificial intelligence "crashed the party" at this year's Cannes Film Festival, with AI tools, AI-generated content, and a high-profile multi-year partnership between Meta and the festival making headlines. Meta's deal represents one of the most significant formal partnerships between a major AI company and a leading global cultural institution, and signals the film industry's complex, ambivalent embrace of AI — even as screenwriters, directors, and actors unions continue to fight against AI's encroachment on creative labor. The partnership is expected to involve AI tools for content creation, promotion, and audience engagement. The Cannes moment illustrates how AI is rapidly moving from a back-office technology tool into the center of culture and entertainment, forcing creative industries to negotiate terms with tech companies on AI's role in storytelling, rights, and compensation.
- 12
Huawei Unveils 'Tau Scaling Law' as Its Own Answer to Moore's Law, Targeting 1.4nm Chips
Huawei has revealed its own alternative to Moore's Law called the "Tau Scaling Law," which shifts the focus of chip advancement away from pure transistor shrinking and toward system-level architectural efficiency. The move positions Huawei as it aims to develop 1.4nm-class chips domestically, circumventing US export restrictions that have cut off its access to TSMC and advanced Western chip manufacturing. The Tau Scaling Law is designed to guide Huawei's chip roadmap in a post-Moore era, emphasizing innovations in chip packaging, interconnects, and AI-specific compute designs. The announcement signals China's growing determination to develop an indigenous semiconductor ecosystem independent of Western supply chains, and could reshape how the global industry measures and tracks chip progress — particularly as Nvidia and other US chipmakers push their own next-generation architectures.
- 13
People Are Deliberately Adding Typos to Signal Their Writing Is Human, Not AI
Futurism reports on a growing cultural trend in which people are deliberately seeding their writing with typos, grammatical quirks, and imperfections as a way to signal that their content was written by a human rather than generated by AI. As AI-produced text becomes increasingly polished and indistinguishable from human writing, the imperfection paradox has emerged: audiences now associate perfect grammar and flawless prose with AI, while human-sounding errors convey authenticity. The trend is appearing across social media, professional emails, creative writing communities, and even journalism, as the cultural anxieties around AI-generated content reach everyday communication. The phenomenon reflects a deeper identity crisis in the age of AI — a society recalibrating what "human" means in a world where machines can imitate it flawlessly, and where authenticity has become a scarce and deliberately performed commodity.
- 14
Copilot Gets Dockable Sidebar in Windows 11, Mirroring Gemini's Chrome Integration
Microsoft is adding a dockable sidebar for Copilot in Windows 11, a design move that closely mirrors how Google has integrated Gemini as a persistent sidebar within Chrome. The new interface allows users to keep Copilot accessible alongside any application without it taking over the screen, making it a persistent ambient AI layer within the operating system. The update represents Microsoft's latest effort to deepen Copilot's integration into the Windows experience and keep pace with Google's aggressive push to embed Gemini across its ecosystem. The sidebar approach reflects a broader industry trend toward making AI assistants persistent, contextual, and always-on within computing environments — moving AI from an on-demand tool to a constant presence in the user interface. The timing is notable as competition between Microsoft and Google for AI assistant dominance intensifies.
- 15
Record Label Claims AI-Generated 'Viking Rappers' With Millions of Views Are Real People
Futurism reports on a bizarre case in which a record label is claiming that a group of "Viking rappers" accumulating millions of views on social media are real human artists — despite widespread evidence suggesting they are entirely AI-generated personas. The case highlights the accelerating problem of AI-generated fake artists flooding music platforms, and the commercial incentives that drive labels and distributors to obscure the AI origins of content. The fake artist ecosystem has grown significantly as AI music generation tools have become more capable and accessible, with AI-generated tracks routinely gaming streaming algorithms to rack up plays and royalties. The incident underscores urgent questions about transparency, attribution, and authenticity in the AI-generated content economy, and the inadequacy of current platform policies and content verification systems to detect and label AI-generated artists.