Latest AI News

Get the fastest, most accurate AI news every morning

Total articles: 11,274|Added todayToday: 81

Category

Search

Apple Launches HomeView Smart Home Device

Apple launched HomeView, a $299 smart home screen with a 7-inch display and face recognition
Siri handles multiple family members’ voice requests; HomeView requires a recently unlocked iPhone on the same network for private data
Apple added LG and Schneider Electric accessories and priced its new Apple TV 4K at $299

channelnewsasia.comRead more · 2 min →

Today'sSafety & Ethics

Massachusetts Weighs AI Safety Bill

Massachusetts weighs independent AI model testing at least three times a year
Proposed rules set penalties up to $1 million initially and $3 million for later violations
Anthropic and OpenAI support evaluations as lawmakers negotiate the bill’s final provisions

masslive.comRead more · 2 min →

OpenAI’s ChatGPT Space Puts Work in One Place

OpenAI announced more than 20 products and features at DevDay; XDA’s writer singled out ChatGPT Space.
Space offers editable Pages, project organization, search, sharing and real-time collaboration within ChatGPT.
ChatGPT Space and Dots require Pro, starting at $100 per month; Plus costs $20 monthly.

XDA DevelopersRead more · 1 min →

Today'sSafety & Ethics

SpaceX Agrees to NYC Council AI Safety Testimony

SpaceX agrees to testify this week in New York City Council’s AI safety investigation
Council says SpaceX missed the Oct. 5 hearing attended by Anthropic, OpenAI, Google and Meta
Subpoena enforcement threat preceded agreement; investigation cites privacy, cybersecurity and public safety risks

CNBCRead more · 1 min →

White House Task Force to Press AI Companies on Safety

White House task force to coordinate with Justice Department and law enforcement on AI safety expectations
Vice chair Scott Kupor says standards will draw on existing criminal and civil laws
Group plans to audit AI accord pledges, assess risks and prepare government response resources

PoliticoRead more · 1 min →

Today'sSecurity

OpenAI Finds Hidden Recording Device in Office

OpenAI reported finding a sophisticated recording device hidden inside one of its offices
The company said its models helped identify an employee who planted it days earlier
OpenAI referred the case to the FBI; it says the investigation is ongoing

twitter.comRead more · 1 min →

Review Finds Limits in Clinical AI Safety

A 2026 narrative review assesses generative AI’s effects on clinical decisions and patient safety
GPT-4 assistance raised physicians’ management reasoning scores by 6.5 percentage points
Only about 5% of evaluations used real patient-care data; review recommends supervised clinical use

Semantic ScholarRead more · 1 min →

Trending Keywords

Plano Approves $390,000 Citywide AI Initiative

Plano City Council approved up to $390,000 for citywide AI tools and governance.
ChatGPT Enterprise starts with 150 licenses costing an estimated $57,024 annually.
Darwin AI, Teams Premium and reserve costs account for $200,000 of the package.

GovTech AIRead more · 1 min →

Study Projects $16B in Ohio Power Costs

Scioto Analysis projects $16 billion in additional Ohio electricity costs from 2027 through 2034 under existing policies.
The estimate applies a national study’s 16% middle-demand scenario to Ohio retail electricity sales data.
The report extends modeled increases through 2034 and omits a complete year-by-year calculation for independent reproduction.

GovTech AIRead more · 2 min →

Nutanix Says AI Infrastructure Costs Fell 50%

At its Seoul conference, Nutanix outlined ways to reduce AI infrastructure costs and operational complexity.
After a vendor pricing change drove token costs up 350% overnight, Nutanix shifted to buying its own GPUs.
Nutanix says hybrid infrastructure cut operating costs 50%; it plans an agent gateway to improve control of token use and costs.

Byline NetworkRead more · 2 min →

Colorado Seeks to Normalize AI Across Government

Colorado’s first AI architecture director is shaping the state’s approach during an OIT realignment
Jane Yang’s three pillars are accelerating delivery, activating opportunities and maintaining transparency
Colorado plans to sunset its central AI team within three to five years as AI work spreads across agencies

GovTech AIRead more · 1 min →

T-Mobile Uses AI for Network Resilience

T-Mobile applies AI to network adjustments during disasters and high-demand periods, according to CTO John Saw.
The carrier made 30,000 antenna adjustments with intent-based automation during January's winter storm Fern.
T-Mobile's backup network assets include hybrid generators, satellite links, mobile cell sites and SuperBroadband.

InformationWeekRead more · 1 min →

Today'sSafety & Ethics

IT Executives Brace for AI Agent Incidents

Workato survey finds executives expect AI agents to cause public security or compliance incidents within 18 months
36% would handle an agent-driven security incident quietly; 1,001 IT decision-makers surveyed
42% seek visibility into agent workflows as enterprises begin centralizing AI agent governance

CIO DiveRead more · 1 min →

Arena Releases Max Agent Router

Arena released Max, an Agent Arena router reporting frontier performance with 37% cost savings against top models.
In a live test, Max ranked #1 at +14.7%, costing a median $3.27 per task.
Max routed code, chat and work sessions among Astra and Fable 5.1 models at different rates.

Arena AIRead more · 1 min →

Workday Adds Autopilot Agents for Finance

Workday announced autopilot agents for finance and human resources workflows on Tuesday, Oct. 13, 2026
Payroll and accounting agents are planned for early adopters in early 2027
Customers can set automated actions, reviews, final decisions, and monitor logged activity and costs

CIO DiveRead more · 1 min →

Higgsfield Expands AI Video Beyond 10 Minutes

Higgsfield says GPT-6 Astra enables videos longer than 10 minutes, up from around one minute.
The company uses GPT-6 Astra for instructional video production and creative agents’ 3D modeling tasks.
Higgsfield uses GPT-Image-2.5 and Codex, and aims to let creators shape films through spoken ideas.

OpenAIRead more · 1 min →

Grok 4.7 Arrives on Amazon Bedrock

Amazon Bedrock adds SpaceXAI’s Grok 4.7 with configurable reasoning effort for tool-calling agents
AWS test: 67 correct answers in 72 runs; xhigh scored 18 of 18, low scored 15
Higher effort used more output tokens; AWS recommends repeated task-specific testing and production monitoring

AWS ML BlogRead more · 2 min →

AWS Reports MoE Inference Optimizations

AWS details three fused MoE kernel optimizations for Amazon SageMaker AI inference
Operand swapping improved output throughput by 2–4 percent on H200; Triton routing raised throughput 10 percent
Triton recompilation elimination improved TTFT by 9 percent at 8-way tensor parallelism on H200

AWS ML BlogRead more · 1 min →

AWS Adds Managed Payments for Browser Agents

AWS lists Anchor Browser and AgentCore payments integration on AWS Marketplace for agents accessing x402-protected pages and APIs.
AgentCore checks session budgets and signs payments through Coinbase CDP or Stripe Privy without exposing wallet keys to agent runtimes.
Developers enable payments with session.x402=true; existing browser automation can retry paid requests while session spending limits apply.

AWS ML BlogRead more · 2 min →

MIT Network Builds Its Own Learning Machinery

MIT researchers report a self-organizing artificial network that builds learning machinery from simple connection rules
The network recognized patterns and classified images, performing at a similar level to back-propagation networks
The team says self-organization offers a starting point for studying biological learning and brain-machine connections

MIT AI NewsRead more · 2 min →

Trending Keywords

Last 7 Days