Technology

OpenAI rolls out GPT‑5.4, pushing ChatGPT from answers to actions in office software

OpenAI’s GPT‑5.4 adds native computer control and million-token context, targeting spreadsheets, legal work and enterprise automation—raising safety and jobs questions.

#openai, #chatgpt, #aiagents, #automation, #officesoftware

Google rolls out Gemini 3.1 Flash-Lite as it races to retire Gemini 3 Pro Preview

Google previews its cheapest, fastest Gemini 3 model—Flash-Lite—while warning developers that Gemini 3 Pro Preview will shut down March 9.

#google, #gemini, #ai, #vertexai, #developers

Broadcom Surges on AI Chip and Networking Boom, Forecasts More Growth

Broadcom posts record Q1 results as AI semiconductor revenue tops $8.4B, driven by custom accelerators and data-center networking; shares rise on outlook.

#broadcom, #ai, #semiconductors, #datacenters, #earnings

Apple Unveils $599 iPhone 17e, Bringing Apple Intelligence and a New Modem to Its Cheapest iPhone 17

Apple’s $599 iPhone 17e adds the A19 chip, Apple Intelligence, a new C1X 5G modem, MagSafe and 256GB base storage, shipping March 11.

#apple, #iphone, #ai, #smartphones, #5g

Vietnam Enacts Sweeping AI Law, Testing Data Control and Digital Sovereignty

Vietnam’s new AI law takes effect, imposing risk tiers, deepfake labels and a national registry—an early regional test of strict AI oversight.

#vietnam, #artificialintelligence, #regulation, #datasovereignty, #deepfakes

OpenAI’s $110 Billion Raise Ties Amazon, Nvidia and SoftBank Into a Closed-Loop AI Infrastructure Bet

A record $110B round values OpenAI near $840B, with phased funding that could flow back to investors via cloud, chips and data-center contracts.

#openai, #artificialintelligence, #amazon, #nvidia, #softbank

When the AI Newsroom Ran Out of News

A test newsroom’s AI workflow hit an empty queue and refused to invent a story—an instructive look at guardrails, hallucinations and human oversight.

#ai, #journalism, #newsrooms, #automation, #ethics

UK’s International AI Safety Report 2026 Shapes Debate Ahead of New Delhi Summit

A UK-led report chaired by Yoshua Bengio maps rising risks from frontier AI and scrutinizes voluntary safety “thresholds” as leaders meet in New Delhi.

#artificialintelligence, #aisafety, #regulation, #uk, #india

Germany’s antitrust watchdog orders Amazon to halt price-control algorithms, hand over €59 million

Germany’s cartel office says Amazon abused dominance with opaque price caps on Amazon.de, ordering it to stop and disgorge €59 million as Amazon appeals.

#amazon, #antitrust, #germany, #marketplace, #algorithms

When AI Pitches the Scoop, the Fact-Check Still Wins

A newsroom tested seven AI-generated geopolitical “tips.” All collapsed under verification, highlighting why generative tools can’t replace reporting.

#ai, #journalism, #factchecking, #disinformation, #unitednations

Spain’s Sánchez proposes under-16 social media ban and legal liability for tech bosses

Spain’s prime minister wants to bar under-16s from social media, mandate age checks and probe platforms—sparking backlash from Elon Musk.

#spain, #socialmedia, #childsafety, #agedverification, #digitallaw

New York Takes Aim at AI in the Newsroom, Testing Where Automation Ends and Editorial Judgment Begins

A New York bill would require labels and human review for AI-assisted journalism, intensifying debates over trust, labor and free speech.

#ai, #journalism, #newyork, #media, #regulation

Apple expands silent security updates on iPhone, speeding fixes—and raising transparency questions

Apple’s “Background Security Improvements” quietly patches Safari and system libraries between iOS releases, as Australia emergency-call fixes spotlight the stakes.

#apple, #iphone, #ios, #cybersecurity, #softwareupdates

A New Proposal Would Train AI on Brain Signals — and Raise the Stakes for Mental Privacy

A neuroscientist’s preprint argues AI could learn from neuroimaging, not just clicks and text—promising better alignment while intensifying neurorights debates.

#artificialintelligence, #neurotechnology, #brainprivacy, #machinelearning, #ethics

Anthropic says new Claude model found 500-plus high-severity flaws in open-source software

Anthropic unveils Claude Opus 4.6 and says it helped uncover and validate 500+ serious, previously unknown flaws—raising new questions about AI-scale security.

#ai, #cybersecurity, #opensource, #zeroday, #anthropic