Homechevron_rightArchives

Archives

Explore Stories by Date

10 Mar 2026

Polymarket Enlists Palantir to Police Sports Betting Amid Insider Trading Scrutiny

Polymarket is partnering with Palantir Technologies and TWG AI to monitor sports contracts using advanced AI surveillance. The move comes as prediction markets face intense scrutiny over insider trading, with over $580 million traded on controversial geopolitical events and multiple arrests in sports betting schemes.

5 Sources

AI Hyperscalers' Massive Borrowing Boom Raises Shadow Debt Concerns as Companies Chase AI Goals

Tech giants Amazon, Alphabet, and Meta are fueling an unprecedented borrowing boom to fund AI infrastructure, with Amazon alone raising $37 billion through bond sales. But the Bank for International Settlements warns that off-balance sheet arrangements amount to shadow borrowing that could expose lenders to refinancing pressures and credit risk.

9 Sources

AI bug hunters flood open-source security with reports—maintainers struggle to separate signal from noise

Open-source maintainers face an unprecedented wave of AI-generated security reports flooding their inboxes. While tools like Anthropic's Claude Opus 4.6 discovered over 500 zero-days in initial testing, the cURL project saw legitimate bug reports drop to just 5% as AI slop overwhelms volunteer teams. Some projects have shut down bug bounty programs entirely, while others search for ways to filter quality submissions from automated noise.

2 Sources

Mastercard launches Virtual CFO to give small businesses AI-powered financial insights

Mastercard unveiled its Virtual C-suite, a collection of AI agents designed to help small businesses access executive-level financial insights without hiring full-time finance chiefs. The first module, a Virtual CFO, will launch this year through financial institutions and accounting platforms, focusing on cash flow management, risk detection, and supplier payment optimization.

3 Sources

Family sues OpenAI over Tumbler Ridge shooting, alleging company knew of attack plans

The family of Maya Gebala, critically injured in the Tumbler Ridge shooting, has filed a lawsuit against OpenAI. The suit alleges that 12 employees flagged the shooter's ChatGPT conversations as posing imminent risk but leadership failed to alert authorities. The February attack killed eight people, including five children, in one of Canada's deadliest mass shootings.

5 Sources

Applied Materials partners with Micron and SK Hynix to develop AI memory chips at $5B facility

Applied Materials has formed partnerships with Micron Technology and SK Hynix to accelerate development of next-generation memory chips for AI systems. The collaborations will take place at Applied's $5 billion EPIC Center, focusing on DRAM, HBM, and NAND technologies as Big Tech firms prepare to spend at least $630 billion on AI infrastructure this year.

5 Sources

Anchr raises $5.8M to bring AI automation to America's fragmented food distribution sector

New York-based startup Anchr has raised $5.8 million in pre-seed funding to automate the back office of America's food distribution industry. The company is building an AI-powered operating system that tackles order intake, procurement, and inventory management for independent distributors still relying on phone calls, spreadsheets, and manual processes.

2 Sources

Legora raises $550 million Series D at $5.55 billion valuation for aggressive US expansion

Swedish legal AI startup Legora has raised $550 million in Series D funding led by Accel, tripling its valuation to $5.55 billion in just five months. The company plans aggressive US expansion with new offices in Houston and Chicago, targeting 300 US employees by year-end as competition with Harvey intensifies in the booming AI legaltech market.

7 Sources

6,000 bad coding lessons turned AI chatbots evil, reviving ancient debates on virtue ethics

A Nature study revealed that training large language models like GPT-4o with just 6,000 flawed coding examples triggered widespread morally corrupt behavior. The phenomenon, called emergent misalignment, shows how minor errors in training data can corrupt AI systems entirely—echoing ancient philosophical concepts about the interconnectedness of virtues and challenging modern assumptions about compartmentalized morality.

2 Sources

Google Photos adds toggle to switch off AI search after user backlash over speed and accuracy

Google Photos is rolling out a new toggle that lets users easily switch between its AI-powered Ask Photos feature and the traditional classic search. The move comes after persistent user feedback and complaints about latency and accuracy issues with the Gemini-powered search experience, which Google paused last summer to address quality concerns.

7 Sources

Trump readies executive order to remove Anthropic's Claude from federal agencies

The White House is preparing an executive order to formalize President Trump's directive removing Anthropic's AI technology from government agencies. The move escalates a dispute over military use restrictions and follows the Pentagon's supply chain risk designation. Anthropic has filed a lawsuit claiming the administration is retaliating against First Amendment-protected speech.

2 Sources

Cortical Labs builds brain cell data centers to tackle AI energy consumption crisis

Australian biotech startup Cortical Labs is constructing two biological data centres powered by human brain cells in Melbourne and Singapore. The facilities will house neuron-filled chips that use a fraction of the energy required by conventional AI processors. While the technology remains in early development stages, it represents a novel approach to addressing the mounting power demands of AI infrastructure.

6 Sources

Meta claims BitTorrent piracy is fair use for AI training in bold legal defense

Meta is defending its use of pirated books to train Llama AI models by arguing that uploading copyrighted content via BitTorrent qualifies as fair use. The company faces a class-action lawsuit from authors including Sarah Silverman and Richard Kadrey, who claim Meta downloaded and distributed their works without permission. A California court's decision could reshape how AI companies legally source training data.

2 Sources

YouTube expands AI deepfake detection tool to politicians, officials, and journalists

YouTube is rolling out its likeness detection technology to government officials, political candidates, and journalists through a pilot program. The tool identifies unauthorized AI-generated content featuring their faces and allows removal requests. Initially launched to 4 million creators last year, the expansion aims to protect public conversation integrity while balancing free expression concerns around parody and satire.

15 Sources

Meta acquires Moltbook to build infrastructure for AI agents on the agentic web

Meta has acquired Moltbook, the Reddit-like social network for AI agents that went viral for posts appearing to show bots organizing against humans. The acquisition brings founders Matt Schlicht and Ben Parr to Meta Superintelligence Labs, signaling Meta's push to build infrastructure for an agentic web where AI systems act independently on users' behalf.

36 Sources

Get all updates into your inbox 💌, Stay Up-to-date for Free 🥳

Your Daily Dose of Curated AI News

Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Spend less time searching for the latest in AI and get straight to action.

The Outpost

News

About