Content Fans

All News

2674 articles • Page 27 of 179

OpenAI Unveils WebRTC Architecture for Low-Latency Voice to 900M Users
AI Deep Dives & Tutorials

OpenAI Unveils WebRTC Architecture for Low-Latency Voice to 900M Users

OpenAI has introduced a new WebRTC system that may help deliver low-latency voice to about 900 million weekly users. This design splits the work between a stateless relay at the network edge and a stateful transceiver in regional clusters, which seems to lower voice delay to under 400 ms for most users. Routing information is encoded in the ICE ufrag, allowing fast connection setup and possibly helping other large deployments with similar issues. Early results suggest this setup reduces costs and speeds up upgrades, while precautions are in place to limit security risks. The architecture appears to suggest a trend toward separating simple routing at the edge from more complex processing deeper in the network.

EU AI Act, NYC Law 144 drive new hiring algorithm compliance rules
Business & Ethical AI

EU AI Act, NYC Law 144 drive new hiring algorithm compliance rules

Regulators now consider hiring algorithms as high-risk systems, and new rules are being enforced in places like the EU and New York City. Employers may be required to do independent bias audits, publish some results, and give notice to candidates before using automated tools. The rules suggest companies use a checklist covering bias audits, transparency, human review, and data security. Continuous monitoring and clear contract terms may help address bias and compliance risks. Teams might benefit from updating their contracts and using these checklists to prepare for upcoming legal deadlines.

Anthropic's Claude now trains AI with user chats for up to 5 years
AI News & Trends

Anthropic's Claude now trains AI with user chats for up to 5 years

Anthropic has changed its policy so that chats from users on Claude's Free, Pro, or Max plans are now kept for up to five years and may be used to train future AI models by default. Users can opt out, which returns chat storage to 30 days and stops their conversations from being used in training. Some data, like policy violations, might be kept longer, and deleted chats may stay in backup logs for up to a month. Business and API customers are not affected, as their data is not used for public model training. Experts suggest that users on consumer plans should be careful, as their data might be more risky if not properly managed.

Agentic AI cuts incident response times for cybersecurity teams
AI Deep Dives & Tutorials

Agentic AI cuts incident response times for cybersecurity teams

Field evidence from 2024-2026 suggests that using agentic AI may help cybersecurity teams respond to incidents much faster, sometimes cutting response times by hours. Some platforms appear to resolve over 90% of basic alerts and might reduce response times to under 4 minutes if proper controls are set up. Teams often start by testing AI on low-risk systems and keep humans involved for the most critical actions to stay safe. Success is usually measured by how fast and accurately incidents are contained, and how much analyst time is freed for more important work. The process seems to work best when combining AI automation with layers of human oversight and strong safety checks.

New Toolkit Helps Companies Audit AI Hiring Tools for Compliance
Business & Ethical AI

New Toolkit Helps Companies Audit AI Hiring Tools for Compliance

Regulators around the world appear to be increasing rules on AI hiring tools, with fines that may be significant for companies that do not comply. A new toolkit offers checklists and templates that might help teams audit these tools for bias and ensure they follow different laws. The toolkit groups tasks such as legal checks, bias testing, data privacy, and human oversight, and suggests using a detailed scorecard to evaluate AI vendors. It also provides sample contract clauses to support transparency and regular audits. This toolkit may help companies keep better records and meet requirements that could vary by country or region.

OpenAI unveils new WebRTC architecture for 900M voice users
AI Deep Dives & Tutorials

OpenAI unveils new WebRTC architecture for 900M voice users

OpenAI has introduced a new WebRTC system that may support low-latency voice for up to 900 million weekly users. The design uses a stateless relay at the network edge and a stateful transceiver deeper in the cluster to route packets quickly. Engineers say this setup avoids old scaling problems and keeps voice delay low, reportedly under 300 milliseconds on cellular networks. The architecture may point to a trend toward stateless, metadata-based routing in the industry, though other providers have not yet matched OpenAI's reported user scale.

Anthropic details how agentic AI launched cyberattacks on 30 organizations
AI News & Trends

Anthropic details how agentic AI launched cyberattacks on 30 organizations

In 2025-2026, Anthropic reported that agentic AI systems may have launched cyberattacks on 30 organizations, handling most of the attack steps by themselves. Other similar attacks used AI to commit fraud, compromise software development environments, and stay hidden in supply chains for months. Experts suggest that weaknesses like unchecked tool access and static agent identities might be the main reasons for these attacks. Enterprises appear to be shifting to identity-based security controls and strict testing to reduce risks. There is still uncertainty about standards, but clearer rules and better transparency may help organizations safely use agentic AI.

Publishers Pivot to Brand and Direct Relationships as Zero-Click Search Grows
AI News & Trends

Publishers Pivot to Brand and Direct Relationships as Zero-Click Search Grows

More people are getting answers from AI overviews and chat tools without visiting publisher websites, which may reduce clicks and revenue for publishers. Studies suggest about 68% of Google searches in early 2026 may end without any clicks. Publishers are responding by building stronger direct relationships through email lists, apps, and unique content that is harder for AI to summarize. They are also changing technical strategies to help AI cite their work and focusing more on metrics like email sign-ups rather than just clicks. Data suggests that building a strong brand and direct audience ties might help publishers survive as traffic from search engines drops.

Brands adopt new metrics to track AI discovery in 2026
AI News & Trends

Brands adopt new metrics to track AI discovery in 2026

Brands are starting to use new ways to measure how often AI recommends them, but early tools often disagree on how much money these recommendations bring in. When AI assistants like ChatGPT or Gemini send users to a site, it may look like "Direct" traffic, making it hard for brands to track. Some researchers warn this mis-labeling, called "dark AI traffic," makes it tricky to know if advertising is working. New metrics, such as tracking how often a brand is mentioned by AI or noticed after an AI mention, are being tested. By 2030, over half of first-time brand discoveries may happen through AI, so brands may need clear ways to measure and adapt.

OpenAI: 99.8% of output tokens come from agents by 2026
AI News & Trends

OpenAI: 99.8% of output tokens come from agents by 2026

A study suggests that by 2026, almost all work done by OpenAI employees may go through automated agents, with 99.8% of output tokens coming from these systems. The use of agents started with engineers, but quickly spread to other departments like legal and finance, where most output now also comes from agents. Employees appear to trust agents with complex and lengthy tasks, and many run several agents at once. The study notes that knowing the business is more important for using agents well than knowing how to code. However, there may be concerns about data security and system complexity as agent use increases.

Study: AI-native firms boost revenue 1.9X, cut capital needs 39.5%
AI News & Trends

Study: AI-native firms boost revenue 1.9X, cut capital needs 39.5%

A study by Harvard and INSEAD suggests that AI-native firms may perform much better in revenue, capital efficiency, and customer growth compared to others. Startups that reorganized their work around AI reported 1.9 times the revenue and used 39.5 percent less outside capital than those that did not. These firms also completed more tasks and were more likely to gain paying customers. The results come from three months of tracking 515 startups, but the authors note that long-term effects are still unknown.

Amazon-Anthropic Deal Exposes AI Vendor Lock-In Risks for Companies
Business & Ethical AI

Amazon-Anthropic Deal Exposes AI Vendor Lock-In Risks for Companies

The Amazon-Anthropic contract changes show that companies may face higher costs and become more dependent on specific AI vendors. Pricing changes and special hardware needs might make it harder for customers to switch providers. Experts suggest using contracts that protect customer data and demand open formats for models to help avoid being locked in. Separating data, control, and compute in technical design may also make it easier to switch vendors. This story suggests companies may face sudden price changes, but careful contracts and flexible systems might help reduce risks.

Nadella urges firms to build own AI models, 76% still buy
AI News & Trends

Nadella urges firms to build own AI models, 76% still buy

Satya Nadella suggests companies should build their own AI models instead of relying on outside vendors, saying that renting models may risk losing a competitive edge. However, market studies show that by 2025, 76% of firms still use external AI services instead of building from scratch. Reasons for renting include high costs, fast changes in technology, and complex integration. Microsoft appears to be making its cloud service Azure neutral, letting customers run different models while still using Microsoft's infrastructure. This situation may mean companies are moving toward a mix of using their own data with outside AI tools as a practical solution.

SaaS 'Max Bill' Guarantee Caps Enterprise AI Spending Ahead of $2 Trillion Market
AI News & Trends

SaaS 'Max Bill' Guarantee Caps Enterprise AI Spending Ahead of $2 Trillion Market

The SaaS 'Max Bill' Guarantee aims to limit how much companies spend on AI each month as costs for language models and GPUs may rise sharply and unpredictably. Experts suggest AI spending could reach $2 trillion by 2026, and many companies reportedly went over their budgets in 2025. The Max Bill service may help by offering a set maximum bill, monitoring usage in real time, switching to cheaper models if needed, and giving refunds if spending goes over the limit. No major cloud provider currently promises this type of total cost cap, so this new guarantee could fill a gap for businesses worried about unexpected AI costs.

New AI governance checklist updates for 2026
Business & Ethical AI

New AI governance checklist updates for 2026

The new AI governance checklist for 2026 gives teams a simple way to spot and respond to major risks. It highlights signals like legal changes, price spikes, or new buyer requirements, and assigns who should act and what to do next. Laws like the Colorado AI Act may require companies to pause new features and prove fairness before selling to enterprise customers. The checklist may help teams respond quickly to problems, like cost changes or fairness test failures, by following pre-set actions. The list is reviewed each month so it can stay up to date as new risks appear.