Evolving AI Agents & Ads: Latest from OpenAI and Meta
Tak@Here are today's top AI & Tech news picks, curated with professional analysis.
Decisions | OpenAI API - OpenAI Developers
Expert Analysis
OpenAI has launched the "Decisions API" in public beta, designed to evaluate text and images and return structured, typed answers for applications at approximately 10 times the speed of the traditional Responses API. This API is intended for tasks such as content classification, request routing, and work prioritization.
Currently, the Decisions API exclusively supports the gpt-6-luna model and offers three primary question types. These include "predicate" for estimating the probability of a condition being true, "choice" for selecting one option from a predefined set, and "score" for rating an input against ordered levels based on a rubric.
Practical applications demonstrated include checking product photos for visible damage, routing customer complaints to the correct department, and assessing the severity of an issue. The API's ability to combine multimodal inputs (text and images) makes it a powerful tool for automating complex decision-making processes.
- Key Takeaway: OpenAIのDecisions APIは、マルチモーダル入力から高速かつ構造化された意思決定を可能にし、アプリケーションの自動化と効率化を促進する。
- Author: Editorial Staff
OpenAI Decides ChatGPT Would Be Better With More Sponsored Results
Expert Analysis
OpenAI has announced plans to expand the types of advertisements appearing alongside ChatGPT's outputs, introducing a new "visual ad format." These ads will include images to help users better visualize how products and services might fit into their lives.
Initially, these new visual ads will be served alongside images generated within ChatGPT, with expectations for broader integration across the ChatGPT ecosystem in the near future. This expansion is reportedly driven by OpenAI's need to generate significant revenue quickly.
Subscribers to ChatGPT Plus, Pro, or Enterprise plans will be exempt from these advertisements, allowing paid users to maintain an ad-free experience. This move signifies OpenAI's strategy to bolster its advertising business and monetize its free user base.
- Key Takeaway: OpenAIはChatGPTに視覚広告を導入し、収益化を強化するが、有料プランのユーザーは広告から除外される。
- Author: AJ Dellinger
Meta Wants to Write the Industry Standards for Safe Agentic Commerce, Report Says
Expert Analysis
Meta is reportedly collaborating with companies like Walmart, Stripe, and Sierra to establish industry standards for how AI agents should interact with online businesses. This initiative, dubbed a "personal agent protocol," aims to differentiate authorized personal agent purchases from others.
Widespread concerns exist regarding the potential safety and privacy issues that arise from granting AI agents access to user finances and allowing them to operate freely online. A group of major global banks has also called for industry-wide standards, policies, and consumer protections based on principles such as transparency, safety, privacy, choice, and interoperability.
Meta's new personal AI agent, Muse, has faced several security and privacy issues since its launch, leading Amazon to temporarily block Muse agents from its online shopping platform. This standardization effort seeks to mitigate risks like scams and fraud, ultimately fostering wider adoption of AI commerce agents.
- Key Takeaway: Metaは、AIエージェントによる安全なEコマースを確立するため、業界標準「パーソナルエージェントプロトコル」の策定を主導している。
- Author: Ece Yildirim
Follow me!
