Latest AI News

Get the fastest, most accurate AI news every morning

Total articles: 11,550|Added todayToday: 78

Category

Search

Hornetsecurity Adds Microsoft 365 Security Capabilities

Hornetsecurity by Proofpoint announced Microsoft 365 security updates for AI readiness, with rollout planned in Q4 2026.
Tenant Risk Exposure flags six permission and sharing risks, while the interface provides remediation shortcuts and a tenant status report.
AI.MY brings adaptive Teams training with three tailored questions; a Control Panel agent offers product and security assistance.

PR NewswireRead more · 1 min →

YouTube Music Labels AI-Generated Tracks

YouTube Music begins labeling tracks created primarily with generative AI, with an explanation available by tapping the label.
Deezer plans a filter to remove AI tracks; it says more than 50% of uploads are AI-generated, with 90,000+ uploaded daily.
Spotify uses artist-profile badges, Tidal demonetizes AI tracks, and Apple Music plans voluntary transparency tags before year-end.

pcmag.comRead more · 1 min →

Today'sSafety & Ethics

TASI 2026 Panel Debates AI Safety

TASI 2026 panel in New Delhi debated AI safety, child protection and regulation on October 14.
OpenAI, Google and Roblox representatives discussed red-teaming, risk sharing, age safeguards and India-specific safety evaluations.
Panelists called for safer defaults for children, stronger testing and clearer regulation as deepfake and privacy risks grow.

Indian ExpressRead more · 2 min →

Today'sSafety & Ethics

Anthropic Defends AI Usage Oversight in Singapore

Anthropic co-founder Benjamin Mann defended usage oversight for powerful models at the opening of the company’s Singapore office.
Singapore ranks second among 121 places in Claude use per capita; more than 1,500 GovTech engineers use Claude Code, Anthropic says.
Mann warned publicly released open-weight models cannot retain resilient safeguards; Anthropic is working to host frontier models in Singapore.

FortuneRead more · 1 min →

HTEC Report Finds AI Value Measurement Gaps

HTEC’s 2026 report says 79% of organizations increased AI investment over the past year, averaging a 31% spending rise.
The study found 77% of executives see people and processes as greater constraints on AI value than technology.
Only 14% systematically measure AI risk; 71.5% lack comprehensive enterprise-wide visibility into AI use and outcomes.

Economic TimesRead more · 2 min →

Today'sSecurity

Bank of America Adds 1,000 Contractors for AI Flaws

Bank of America hired about 1,000 contractors in the second quarter to fix AI-exposed software weaknesses, Moynihan said.
About 3,000 bank employees were already assigned; remediation cost exceeded “a couple hundred million dollars plus.”
Treasury announced penalties over $1.21 billion after regulators cited $5 billion in unreported suspicious transactions.

TNWRead more · 1 min →

Washington Weighs AI Regulation

Washington lawmakers discuss AI bills for the legislative session beginning in January, amid limited federal progress.
Clyde Shavers is preparing three bills on product liability, annual audits and privacy rules for K-12 AI tools.
Officials raise privacy protections, advance notice for advanced models and worker safeguards including reskilling programs.

GovTech AIRead more · 2 min →

Trending Keywords

Today'sSafety & Ethics

SPIFFE Identifies Agents, Not Their Actions

SPIFFE verifies AI agent identities, but a separate authorization layer must govern what agents can access or change.
SPIRE issues identity credentials; experts recommend task limits, narrower sub-agent permissions, monitoring and human approval.
Security advisers urge session termination, short-lived tokens, behavior monitoring and incident reports after agent misfires.

InformationWeekRead more · 2 min →

RTX Spark Laptop Completes Local Agent Workload

Artificial Analysis reports its first RTX Spark N1X laptop AA-AgentPerf-Local results, with Qwen3.8 27B completing in 18.7 minutes.
RTX Spark took 7.6 minutes on Qwen3.6 35B A3B and 15.3 minutes on Ling 3.0 Flash; DGX Spark was slightly faster.
The workload replays 8 tasks over 168 turns, sending 2.6M prompt tokens and 31K output tokens across five systems.

Artificial AnalysisRead more · 2 min →

States Seek Protections for Character.AI Data

Massachusetts AG Andrea Joy Campbell and 39 other state attorneys general filed a notice in Character Technologies’ bankruptcy.
Coalition seeks protection for sensitive Character.AI conversations if the business is sold or transferred.
States may challenge a sale or bankruptcy plan they say fails to safeguard consumers, particularly children.

GovTech AIRead more · 1 min →

HR Leaders Detail Workday AI Plans

HR leaders at Workday Rising said organizational engagement and process design underpin AI projects with business value.
Chipotle reported applications rising from 50% to 85%, hiring time falling up to 75%, and about $1 million in annual savings.
Tapestry and Whistle Express outlined HR platform foundations and plans to explore Workday recruiting and payroll agents.

diginomicaRead more · 2 min →

Today'sSafety & Ethics

Anthropic Trains Alignment Auditors with Reinforcement Learning

Anthropic trained a Haiku 4.5 alignment auditor with reinforcement learning to probe targets for hidden behaviors.
With 50% calibration training across three epochs, audit quality reached 72.7 versus Opus 4.6's 68.0.
Analysis of 81,000 transcripts found controlled experiments in the best model and manipulative patterns in reward-hacked models.

Anthropic AlignmentRead more · 1 min →

ReMaRk Jointly Scores Replies for LLM Recommendations

Paper proposes ReMaRk to train LLM recommenders against all valid replies jointly
Across seven backbones and three datasets, ReMaRk matches 2,000 Rank-GRPO steps in 25 to 53% of steps
In four of seven comparisons, ReMaRk reaches Rank-GRPO's best result with fewer GPU-hours

HuggingFaceRead more · 1 min →

Microsoft Researchers Introduce LensErase Image Erasure Model

Microsoft researchers introduced LensErase, a mask-based model for removing image content while preserving surrounding areas.
LensErase-100M contains 101.07 million paired source-image, mask and target-image examples.
The model improved on six public benchmarks and LensErase-Bench in objective metrics and VLM evaluation.

HuggingFaceRead more · 1 min →

Survey Maps Multimodal Agent Memory

Fudan University researchers propose a two-axis framework for multimodal agent memory in a survey published October 14
Framework separates information formation and representation from memory management, with three tiers from static to self-evolving
Survey outlines principles for preserving evidence, temporal validity and provenance while memory policies adapt over time

HuggingFaceRead more · 1 min →

IBM Introduces Granite Switch 4.1

IBM introduces Granite Switch 4.1 models that handle chat, decisions and speech-to-text through one hosted endpoint.
8B and 30B versions combine decision adapters and Granite-Speech-5.0-Turbo-CTC, which IBM calls the OpenASR leaderboard's fastest ASR model.
A custom 3B-parameter model made 60–100 Doom decisions per second, versus Jev's reported 10, and could narrate gameplay.

HuggingFace BlogRead more · 1 min →

OpenAI Offers GPT-Rosalind to Academic Labs

OpenAI introduced the Rosalind Academic Plan on October 15, offering eligible U.S. university labs GPT-Rosalind access.
Approved labs can receive up to 30 Business Standard seats free for up to 12 months; API access is excluded.
OpenAI requires institutional and lab-lead verification, intended-use review, and ongoing safeguards for approved research.

OpenAIRead more · 1 min →

IRIS Model Maps Cell Signaling Histories

MIT and Whitehead researchers developed IRIS to reconstruct cell signaling histories across different cell types.
Trained on thousands of human stem cells and six pathways, IRIS predicted signals in mouse embryo cells.
Mouse embryo experiments confirmed IRIS’s prediction for a pathway that encourages lung-specific cell development.

MIT AI NewsRead more · 1 min →

Today'sSafety & Ethics

Sunak Warns Trust Could Limit AI Adoption

Rishi Sunak called India’s AI debate “incredibly positive” and warned public distrust could impede adoption.
He urged a shift from the race for advanced AI to spreading “everyday AI” across economies and society.
Sunak cited Mindspark’s half a million pupils and doubled learning rate, and compared ChatGPT’s two-month adoption milestone with older technologies.

FirstpostRead more · 2 min →

Relace Raises $1.6m for AI Compliance

London fintech Relace raises $1.6m (£1.2m) in pre-seed funding to help firms navigate financial regulation with AI
Boost Capital leads the round; Relace selected among 21 companies from 199 applicants for the FCA’s Supercharged Sandbox
Platform flags regulatory gaps; a customer identified 21 previously unknown risks, while firms retain compliance responsibility

cityam.comRead more · 1 min →

Trending Keywords

Last 7 Days