PromptsZone Logo

PromptsZone

AI news & analysis

AI News

Reported developments in AI models, products, research, and the industry.

Conceptual glass archive with one open section, one frosted section and a magnifying glass, illustrating independent scrutiny with access limits.

OpenAI opens embedded safety assessments, with access and disclosure limits in focus

November 17th

OpenAI names four embedded assessors. Apollo’s pilot explains why access, public verdicts and the ability to disclose limitations are separate safeguards.

Read more
AI News
Study respondents’ reasons for expecting faster scientific discovery: faster execution 52%, more promising questions 23%, wider range 14%, other 11%; 367 reasons from 224 responses.

OpenAI expands researcher grants, as pilot study separates time savings from discovery

November 16th

OpenAI’s next academic-researcher grants open November 23 with ChatGPT Plus. Read the access changes and what the companion pilot study measured.

Read more
AI News
Conceptual paper illustration of healthcare information moving between systems toward a human review checkpoint; not a product screenshot.

ChatGPT for Healthcare adds agents and plugins, with review controls in focus

November 16th

OpenAI’s November 16 healthcare update adds computer-use agents and plugins. What is available, what the evaluation figures mean, and where human review belongs.

Read more
AI News
Conceptual illustration of coarse and fine square grids on a desk

OpenAI launches science blog with a candidate Standard Model construction

November 15th

OpenAI’s new science blog opens with Nathaniel Craig’s GPT-6 Astra-assisted lattice research. What is new, and what remains unresolved.

Read more
AI News
Conceptual illustration of three wooden bridge candidates, with one beneath an independent testing rig; not Google equipment or an actual experiment.

Google’s Antigravity research update puts verification at the center of agent teams

November 14th

Google’s November 13 Antigravity update highlights agent teams and independent checks. Explore the workflow, current access and what remains forthcoming.

Read more
AI News
Conceptual paper illustration of a human approval stamp between a wallet and a shopping parcel.

ChatGPT Wallet rolls out to U.S. Pro users, with approval required for each purchase

November 14th

ChatGPT Wallet begins its U.S. Pro rollout with per-purchase approval, substitute card credentials and practical checks for merchant checkout.

Read more
AI News
Original chart: decision models represent 26% of October 2026 Vercel AI Gateway requests and 0.53% of estimated spending, each on a separate 100% scale.

Vercel’s AI traffic report shows why request share is not spending share

November 13th

Decision models drew 26% of Vercel AI Gateway requests but 0.53% of estimated spending in October. The distinction matters for AI budgets.

Read more
AI News
Microsoft-commissioned IDC survey: 66% report substantial or transformative AI results; 48% report scaling value across multiple business units. Separate survey measures, not a conversion funnel.

Microsoft’s IDC survey finds an AI scaling gap: what the return figures show

November 13th

Microsoft’s IDC survey separates reported AI results from business-wide scale. Read the return figures, sponsorship and practical measurement limits.

Read more
AI News
Conceptual illustration of child- and adult-sized chairs beside speech bubbles examined through a magnifying lens, representing research on children’s interactions with AI.

OpenAI pledges $2 million for child AI safety research: what the charter promises

November 12th

OpenAI pledges $2 million for research on AI and children. The charter separates industry funding from independent conclusions and requires unfavorable findings to be published.

Read more
AI News
Conceptual illustration of medical and coverage documents passing through a reading lens into an evidence packet for human review, with the official Gemini family logo.

Gemini for healthcare enters U.S. preview, with read-only EHR access

November 11th

Google's healthcare agent enters U.S. preview. Check eligibility, read-only EHR connections and the record limits that matter before deployment.

Read more
AI News
Concept illustration of a miniature storefront connected to product information and conversation symbols, with the official OpenAI wordmark.

OpenAI launches Businesses in ChatGPT, with separate rules for profiles and agents

November 10th

Businesses in ChatGPT enters beta. Understand profile verification, U.S. agent eligibility, current capabilities and what the new insights actually measure.

Read more
AI News
Chart of Anthropic's proposed 825 million-share base IPO: 97.1% company-offered and 2.9% existing-stockholder shares, excluding the underwriter option.

Anthropic launches IPO roadshow with a proposed $70–$78 share price

November 10th

Anthropic proposes 825 million IPO shares at $70–$78. Here's how company shares, stockholder sales and the gross offering value differ.

Read more
AI News
Conceptual ceramic branching pathway with one red branch disconnected inside glass, illustrating isolation rather than actual AI equipment.

NC State researchers test three ways to contain AI backdoors

November 7th

Newly highlighted research targets hidden triggers through model features, architecture and weights, with different requirements and experimental limits.

Read more
AI News
BioBasicsBench mean reward at xhigh effort for seven OpenAI models, with 95% confidence intervals; Astra has the highest point estimate.

OpenAI’s BioBasicsBench tests AI agents on everyday lab evidence

November 7th

A closer look at BioBasicsBench’s developer-reported results and what they mean for supervised lab assistance.

Read more
AI News
Original interval chart showing METR’s 1.9× estimate of Anthropic’s AECI progress-rate acceleration, with a 1.5–2.8× reported 95% confidence interval.

METR finds faster AI development at Anthropic, with major evidence gaps

November 6th

METR estimates faster benchmark progress at Anthropic, but uncertain attribution and redacted evidence limit what the result proves.

Read more
AI News
1/3
1 2 3

Newsletter Subscription

Receive the latest AI news

Blog logo

PromptsZone

AI news and analysis from PromptsZone: model launches, research, AI tools, and the business and policy developments shaping artificial intelligence.

RSS feed

© PromptsZone, 2026.

Made with Blogmaker

×