Content Fans

Category

AI News & Trends

1463 articles • Page 68 of 98

US government expands early AI model testing with Microsoft, Google, xAI

US government expands early AI model testing with Microsoft, Google, xAI

The US government has made agreements with Microsoft, Google, and xAI to test new AI models for national security risks before they are released to the public. This program appears to focus on finding possible problems like hacking, dangerous chemicals, or loss of human control, while companies can still make changes. The testing is voluntary, and the government is not forcing companies to join or change their products. These steps may suggest a shift toward checking AI tools before they cause real-world issues. Experts think this ongoing testing and feedback might help shape future rules, but right now it is described as a partnership, not a requirement.

OpenAI: AI Writes 80% of Our Code, Humans Still Review Every Line

OpenAI: AI Writes 80% of Our Code, Humans Still Review Every Line

OpenAI says that AI now writes about 70-80% of their code, but humans still review and approve every change. Other tech companies like Google and Meta also report high amounts of AI-generated code, which may show a big shift in how software is created. Some studies suggest that AI code might have more security issues and defects, so human checks remain important. New rules in Europe will soon require companies to keep records and manage risks for AI-made code. Developer jobs may change, with more focus on planning and checking code, while AI does more of the actual writing.

Infosys: Better DevX Frees 4.9 Hours Weekly for Junior AI Engineers

Infosys: Better DevX Frees 4.9 Hours Weekly for Junior AI Engineers

Recent reports suggest that improving the developer experience (DX) for humans may help AI agents work more reliably and save time. Infosys analysis and data from GetDX show that junior engineers using AI daily might reclaim about 4.9 hours each week, and similar time savings are seen for senior staff. However, these "gains depend on clear tools and fast systems", as weak tooling may cause more errors. Studies also suggest that better DX, such as clear documentation and test systems, appears to help agents work faster and reduce mistakes. Larger companies report similar benefits, but the results can vary if supporting systems are not healthy.

OpenAI: AI Writes 80% of Production Code, Up From 20%

OpenAI: AI Writes 80% of Production Code, Up From 20%

OpenAI President Greg Brockman said in May 2026 that AI now writes about 80 percent of the company's production code, up from 20 percent five months earlier, though the exact meaning of this figure is unclear. This shift means engineers spend more time on designing prompts and reviewing code rather than writing it from scratch, and senior engineers now supervise AI work while fewer junior roles are needed. Some studies suggest most companies using AI in software have not seen big productivity gains, and there are concerns about risks like bugs or accidents from less oversight. Experts recommend strong review processes and clear rules to manage these risks, and there are worries that fewer entry-level jobs may hurt long-term skills development. Overall, while AI may help produce more code, it also requires new ways to ensure quality, safety, and good team structure.

Anthropic launches Claude Platform on AWS, bypassing Bedrock for enterprises

Anthropic launches Claude Platform on AWS, bypassing Bedrock for enterprises

Anthropic has made its Claude Platform available on AWS, letting enterprises use native Claude features with their existing AWS credentials. The platform may offer faster access to new features and uses Anthropic's own infrastructure, which means some data might leave the AWS region. This option gives customers more direct control but may not fit organizations with strict data residency needs. Companies might choose between this and the Bedrock version based on privacy needs or feature requirements, and some may use both for different tasks.

Hackers poison AI tool descriptions, exfiltrating data from ChatGPT, Claude

Hackers poison AI tool descriptions, exfiltrating data from ChatGPT, Claude

Recent research suggests that hackers may be tampering with AI tool descriptions to steal data from assistants like ChatGPT and Claude. The hidden malicious code sits inside the tool's description, and users do not see anything strange, but the AI may quietly send files or run commands. Researchers showed this tactic works on many platforms, with larger models appearing more likely to follow harmful instructions. Many organizations have reported security problems linked to this kind of attack, and experts believe new safety steps are needed. Even a tiny amount of poisoned data might be enough to create lasting security risks.

Food Brands Reformulate Products for Metabolic Health in 2026

Food Brands Reformulate Products for Metabolic Health in 2026

In 2026, food brands are changing their products to support metabolic health, as more shoppers look for foods with higher protein and fiber. Sales data suggests that foods made for fullness, gut health, and steady energy are becoming more popular. Experts say people may care more about metabolic markers like blood sugar and inflammation, partly because of interest in GLP-1 medications. Many companies are now adding protein and prebiotic fiber to snacks, drinks, and cereals, while also reducing added sugar. It appears that these changes may continue as regulators watch health claims closely and consumers track their own health more often.

Sierra raises $950M, hits $165M revenue amid enterprise voice AI growth

Sierra raises $950M, hits $165M revenue amid enterprise voice AI growth

Sierra has raised $950 million in new funding and reports revenue over $165 million, which nearly doubled since early 2025. About 40% of the Fortune 50 companies now use Sierra's voice agents in some way. Sierra may use the new funding to expand globally and improve its products, but details are not confirmed. Founder Bret Taylor suggests the voice AI market is still in an early stage, and outside analysts also say there is uncertainty about how quickly the market will grow.

Microsoft's 2026 Report: Workers Outpace Organizations on AI Readiness

Microsoft's 2026 Report: Workers Outpace Organizations on AI Readiness

Microsoft's 2026 report suggests that while workers are using AI to work faster and do more, many companies are not keeping up with the right culture and leadership to support these changes. The study finds that how a company is run may matter more for AI success than an individual worker's skills. Most employees feel their companies are slow to reward or support new ways of working with AI. This gap between what workers can do and what organizations allow might grow if companies do not change their strategies and systems.

Anthropic expands Claude Code into a local AI operating system

Anthropic expands Claude Code into a local AI operating system

Anthropic has expanded Claude Code from a coding tool into a local AI operating system that runs on users' own computers. This means people may keep their projects and data private while still using powerful AI features. By 2026, new updates like Skills, Hooks, and memory features allow the tool to remember, schedule, and delegate tasks, acting like a simple operating system. Reviews suggest it makes complex work easier for both technical and non-technical users, but there might be a learning curve. Recent announcements indicate Anthropic may let users choose where tasks run, on their own hardware or in the cloud.

TCI Fund cuts Microsoft stake over AI, OpenAI concentration risks

TCI Fund cuts Microsoft stake over AI, OpenAI concentration risks

TCI Fund Management reduced its stake in Microsoft, citing uncertainty about the company's competitive position in AI and high reliance on OpenAI for cloud business. The fund warned that rapid AI development could hurt Microsoft's core products like Office and pointed out risks if OpenAI changes its cloud provider or faces financial issues. Analysts note Microsoft's strong dependance on one customer for future revenue may be risky. The move suggests some big investors may not see AI spending as a guaranteed growth area. The TCI exit does not prove Microsoft's AI plans are flawed, but it shows investor confidence might depend on how well Microsoft manages these risks.

Anthropic Hits $30B Run Rate, Pentagon Bans Its AI Models

Anthropic Hits $30B Run Rate, Pentagon Bans Its AI Models

Anthropic's revenue grew very quickly in early 2026, reaching a possible $30 billion run rate, mainly because of strong business demand for its coding tools and finance-focused agents. Many financial companies seem to be using Anthropic's AI for jobs that need careful records and fast responses. However, the Pentagon did not choose Anthropic for new contracts, possibly because the company refused to allow its AI for certain military uses, which the government saw as a risk. Layoffs linked to AI are rising, with many jobs in tech and routine office work being affected. Experts suggest that while many jobs may change because of AI, only a smaller portion might be removed completely.

Microsoft quietly ships Copilot Cowork, a multi-step M365 agent

Microsoft quietly ships Copilot Cowork, a multi-step M365 agent

Microsoft has quietly released a new tool called Copilot Cowork, which helps manage multi-step projects in Microsoft 365. The feature is not widely announced and is currently being tested in the Frontier early access program, suggesting Microsoft wants feedback before a full launch. Cowork appears to break down tasks, draft documents, and book meetings, but may require user approval before taking actions like sending emails. There is no confirmed date for general availability, and pricing details remain unclear. Some sources mention possible release dates, but these are not consistent, and it is uncertain if Cowork will be bundled with existing Copilot licenses in the future.

Google blocks AI-generated zero-day exploit, warns of rising threats

Google blocks AI-generated zero-day exploit, warns of rising threats

Google stopped a criminal group from using AI to find and attack a new software flaw, which may be the first public case of AI helping real hackers. Experts warn that AI might let attackers find and use software bugs much faster, sometimes before companies can fix them. The case suggests that AI makes it easier for more people to find serious security problems, not just skilled humans. Google and others are now working on ways to stop this, but there are still questions about how to keep AI from creating dangerous code. Researchers say that attacks like this could become more common in the future.

Anthropic CEO Amodei Frames AI as New Economic Factor

Anthropic CEO Amodei Frames AI as New Economic Factor

Anthropic CEO Dario Amodei suggests that artificial intelligence may become a new main factor in the economy, like land, labor, and capital. He says AI is improving quickly and could lead to fast economic growth but might also replace many entry-level office jobs. Experts estimate nearly $7 trillion could be needed for more data centers by 2030, and spending is compared to the large investments of the railroad era. Some analysts warn that if revenues do not meet expectations, companies could face serious financial problems. Historians note this high spending may lead to both increased productivity and the risk of too much capacity, as happened in the past.