독서리서치레이더투자 프레임워크
로그인 / 회원가입
전체기술제품투자상장시장기타

독서

2026-11-1512 게시물

새로고침

OpenAI Launches ChatGPT Wallet: AI Agents Can Now Shop for Users

중대
2026-11-15 22:56 GMT+8

OpenAI officially launched "ChatGPT Wallet" on Friday, a new payment method designed to enable its AI agents, known as Dots, to execute purchases on behalf of users. This move marks a critical step in transforming AI from an information retrieval tool into an agent capable of actual financial transactions. The feature is currently restricted to ChatGPT Pro subscribers in the United States who are at least 18 years old.

Technically, OpenAI developed this function in collaboration with American Express, Mastercard, and Visa. Users must link their credit card details once, after which they can instruct their AI agent to search for products. Before any charge is made, the system displays item details, amounts, and the selected card for user review, requiring final authorization via biometric methods such as Face ID or fingerprint scanning. OpenAI emphasizes that Dots agents do not access raw card data directly. Instead, they operate using transaction-specific tokens provided by payment networks, which contain temporary card numbers, security codes, and expiration dates. This ensures sensitive information is not exposed to merchants or the agents themselves, mirroring the tokenization approach used by Google Pay or Apple Pay.

OpenAI states that this initiative aims to establish an agent-driven commerce ecosystem, where users might eventually delegate entire shopping processes to AI by simply setting budget limits or timeframes. However, the company acknowledges that many existing websites and payment systems are not yet prepared for such "agentic purchases" and may reject transactions initiated via ChatGPT Wallet. To facilitate this transition, OpenAI released a "Commerce Partner Playbook" to guide businesses in adapting their systems for AI agent compatibility, while also collaborating with fraud prevention specialists like Forter and Riskified, and payment providers such as Adyen and Stripe.

출처:heise online

리서치

70% of UK Workers Uncomfortable with AI Interviews, Yet Nearly Half of Job Seekers Have Experienced Them

주제 · 英国AI面试争议중대
2026-11-15 20:00 GMT+8

According to new research by Careerminds, seven in ten UK workers say they would feel uncomfortable if their interview were conducted with AI assistance, while fewer than a third say they would continue without hesitation. However, the reality contrasts sharply with this hesitation: a separate study by Greenhouse found that as many as 47% of UK job seekers have already been interviewed by AI.

Despite the discomfort, reactions vary widely. About one-third (33%) would continue but feel uncomfortable, and a quarter (25%) say it depends on the role or salary. Only 11% say they would withdraw entirely. Of those, 5.5% would definitely withdraw, and 5.7% are probably considering it. These figures confirm that recruiters could be losing access to vital human talent if they use AI avatars for interviews.

The Greenhouse study put the abandonment figure much higher, at 30% of UK candidates leaving the hiring process due to an AI interview. It stressed that four in five (82%) were not told beforehand that they would be answering questions to an AI avatar. Unsurprisingly, older generations are more likely to withdraw, highlighting challenges for older workers seeking job changes.

Resident careers expert Amanda Augustine noted that interviews serve as a crucial first point of contact, influencing how candidates perceive the employer. She suggested that AI's best fit is helping recruiters with administrative tasks like trawling through applications and refining the candidate pool.

출처:TechRadar

리서치

Reuters: Amazon Plans 'World Product Graph' to Counter AI Shopping Agents

주제 · 亚马逊世界产品图谱중대
2026-11-15 14:03 GMT+8

According to Reuters, US e-commerce giant Amazon is advancing an internal initiative named "World Product Graph," which aims to collect data on nearly every product existing on Earth, regardless of whether it is sold on Amazon's platform. The core intent of this project is to strengthen Amazon's position in the era of AI-powered shopping, directly challenging Google's monopoly in online search and fending off rivals like Walmart.

This strategic move primarily targets the rising trend of AI shopping agents—software capable of autonomously searching and purchasing on behalf of users. Internal documents obtained by Reuters reveal that Amazon's existing AI search agent, Rufus, had provided inaccurate information. Once fully implemented, the World Product Graph is expected to help Amazon retain users who might otherwise leave for other sites due to insufficient data. A 2025 Morgan Stanley study noted that while consumers prefer starting searches on Amazon when they know exactly what to buy (41% vs. Google's 23%), Google dominates broad product research with a 57% usage rate compared to Amazon's 16%.

The project is led by Robert Tekiela, VP of Amazon's Selection and Catalog Systems division. Leaked files indicate Amazon has secured brands like Coca-Cola and Bayer as data contributors and plans to crawl other retailers' websites to supplement its database. However, Amazon officially responded that documents containing future target dates (such as October 2027) are old drafts that do not necessarily reflect current plans. Despite the uncertainty, this effort signals Amazon's attempt to force AI agents to prioritize finding products within its ecosystem by building the most comprehensive product database, thereby protecting its growing advertising business.

출처:The Core(더코어)

리서치

Carnegie Mellon PhD Student Presents 'Tamagotchi-style' AI Agents That Give Everyday Objects Emotions to Nudge Habits

주제 · 卡内基梅隆OATs项目간단한 요점
2026-11-15 19:00 GMT+8

Daehwa Kim, a fifth-year PhD student in the Future Interfaces Group at Carnegie Mellon University in the United States, recently presented a project called OATs (Object Agent Twins) at the ACM Symposium on User Interface Software and Technology (UIST). The project aims to attach AI avatars to everyday physical objects, such as plants, skateboards, or dumbbells, using micro electronic ink (E Ink) displays and sensors, giving them expressive personalities to gently nudge users toward better habits.

The core hardware module of OATs costs approximately $20 and integrates sensors for humidity, temperature, light, motion, and occupancy. The system uses Google Gemini's Nano Banana image generation service to create sprite images for various emotional states. When specific conditions are detected, such as dry soil in a plant or a skateboard that hasn't moved for days, the AI avatar on the screen displays sadness or loneliness; when the user takes action (like watering or exercising), the avatar becomes happy. This design leverages human empathy towards anthropomorphized objects, attempting to provide behavioral prompts without relying on smartphone screens.

In her demonstrations, Kim found that users primarily utilized OATs in three scenarios: confirming if a device is being used (like a coffee machine), warning nearby people of potential dangers (such as toxic fumes from a 3D printer), and tracking personal habits (like fitness frequency). Although the prototype has short battery life (about 3 to 4 days, dropping to 1 day if the sound sensor is active) and is not yet commercialized, it received two Honorable Mention awards at UIST, indicating academic interest in this combination of ambient computing and behavioral intervention.

출처:WIRED

리서치

Australian Security Firm Apate Deploys ~350,000 AI Bots to Counter Phone Scams

주제 · Apate중대
2026-11-15 14:15 GMT+8

Australian security firm Apate has developed an AI-based system to combat the rising number of phone scams. According to German tech media outlet t3n, the system deploys approximately 350,000 AI bots that specifically answer calls from scammers or respond to suspicious messages via instant messaging apps.

The core strategy of these bots is "time-wasting." They mimic various personalities and behavioral patterns, giving scammers false hope that they are about to succeed or that the target is vulnerable, thereby inducing them to invest more time. Dali Kafaar, CEO of Apate, told Wired magazine: "We hope we have built the perfect victims for scammers. Every minute a scammer spends talking to a bot is a minute that potentially protects hundreds, if not thousands, of people from being contacted by the same person."

Beyond consuming scammers' time, these AI agents also infiltrate online scam groups to collect information about criminals and their bank accounts. Kafaar stated that the tools have already gathered over 250,000 pieces of intelligence regarding fraudulent activities and responsible individuals. However, facing the global volume of approximately 68 billion scam calls per year (data from Truecaller), the scale of 350,000 bots remains negligible, and the current primary function is disruption and intelligence gathering rather than completely stopping scam operations.

출처:t3n.de

리서치

Alibaba Launches Standalone iOS and Android App for HappyShrimp AI Music Platform

주제 · 快乐虾米AI音乐平台간단한 요점
2026-11-13 19:05 GMT+8

According to Zhidongxi, Alibaba's AI music creation platform "HappyShrimp" officially launched its standalone APP version on November 13, 2026, simultaneously arriving on iOS and Android platforms. Previously available primarily as a web-based service, this mobile rollout marks a significant step toward broader consumer adoption.

Zhidongxi's hands-on review indicates that the HappyShrimp app allows users to input natural language descriptions of emotions, stories, or musical styles, after which the AI automatically handles lyric writing, composition, arrangement, and vocal performance. The report notes that generating a song approximately three minutes long takes only about one minute. The app reportedly understands complex prompts accurately and can incorporate current events into its creative output, such as generating a farewell song with specific emotional progressions based on news topics.

Beyond original creation, the app supports uploading custom lyrics and specifying reference musical styles for regeneration. The new version also introduces a "Make Same Style" feature, allowing users to discover works in an immersive feed, reuse their creative logic with one click, and substitute personal parameters (such as MBTI personality types) to quickly generate personalized versions. Currently, the platform has generated trending song charts based on user-shared content.

출처:zhidx.com

리서치

OpenAI Launches Daybreak Defense Factory and Red Team for Automated Patching and Testing

중대
2026-11-14 11:13 GMT+8

US-based artificial intelligence company OpenAI officially launched two new products under its Daybreak cybersecurity platform during the 'Cyber at the Frontier' event on November 12. The tools, named Daybreak Defense Factory and Daybreak Red Team, aim to help enterprises accelerate their defense and remediation speeds in response to AI-accelerated attacks. Both services are currently in early trial phases, with no pricing or general availability timeline announced.

Daybreak Defense Factory focuses on automating the patching process. Composed of AI agents built on OpenAI's latest cybersecurity model, GPT-6 Cyber, it scans code, identifies vulnerabilities, and automatically generates fixes. Crucially, these fixes are not deployed directly to production environments; instead, they are submitted to human engineers as 'merge requests' accompanied by validation tests. This design maintains human oversight, ensuring AI-generated patches are confirmed before going live.

Daybreak Red Team specializes in automated red team exercises. The tool uses OpenAI's models to perform continuous scanning and attack path simulation on enterprise web applications. It references source code and application context to investigate potential attack vectors, validates issues reported by other code scanning tools, bug bounty programs, and vulnerability disclosures, and tests the effectiveness of subsequent patches. Nancy Wang, who leads AI applications for cybersecurity at OpenAI, described it as a penetration test at the scale of automated continuous scanning, surpassing human capabilities.

OpenAI President Greg Brockman stated during the event that an enterprise's security level depends on the capabilities of the top-tier models used, requiring cybersecurity practices to be reoriented around what these models can do. He emphasized that while fundamentals remain important, the products themselves are equally critical. For enterprises facing AI-driven threats, the ability to use equivalent-grade tools to shorten the time from vulnerability discovery to remediation is a key metric for evaluating such products.

출처:INSIDE

리서치

DeepSeek Opens Full Suite of Application APIs Including Voice and OCR, Enabling Low-Cost App Replication

주제 · DeepSeek중대
2026-11-14 10:39 GMT+8

Chinese AI company DeepSeek announced on November 14, 2026, that it has officially opened its "full suite" application APIs. Features previously available only within its official App—such as web search, deep thinking, file upload, link reading, and newly added voice recognition (ASR) and speech synthesis (TTS)—are now provided as independent interfaces. Developers can combine these capabilities as needed and are billed separately for each service.

The key update lies in the comprehensive opening of multimodal capabilities. Although DeepSeek launched Search and Fetch interfaces in October, developers previously had to write custom code to orchestrate model decisions, search execution, and content summarization. Now, by simply enabling web search, the model automatically handles search timing, content retrieval, and result integration. Additionally, voice capabilities now have official independent interfaces; previously, developers building voice-enabled assistants typically needed to integrate third-party services. Currently, the V4.1-Flash model supports these extended functions across Messages, Responses, and Chat Completions interfaces.

In terms of pricing, DeepSeek has adopted an aggressive strategy to drive adoption. Voice recognition is discounted by 50% to 0.39 yuan per hour (from 0.78 yuan), and speech synthesis is also halved to 1.5 yuan per 10,000 characters (from 3 yuan). For comparison, OpenAI's GPT-Transcribe costs approximately 1.81 RMB per hour of audio, while Google Cloud's standard speech service reaches 6.43 RMB. Web search and fetching remain at 5 yuan per 1,000 successful calls, and Optical Character Recognition (OCR) is priced at 1.5 yuan per 1,000 pages.

DeepSeek has also open-sourced a lightweight chat frontend Demo that supports importing historical conversations, aiming to help developers quickly build products with functionality close to the DeepSeek App. The company stated that developers can even "replicate the complete DeepSeek App 1:1." This move indicates DeepSeek's attempt to attract developers to build next-generation AI applications on its platform by lowering basic tool invocation costs and providing a complete technology stack, rather than relying solely on its own consumer-facing chat product.

출처:量子位

리서치

South Korean Securities IT Firm DefiSoft Launches Paave 2.0 in Vietnam, Enabling User-Created AI Agents for Simulated Stock Trading

간단한 요점
2026-11-13 16:29 GMT+8

South Korean securities IT firm DefiSoft has officially launched its market analysis and investment platform, "Paave 2.0," in the Vietnamese market. The core innovation of this version allows users to build their own AI agents capable of automatically analyzing real-time quotes, financial statements, and public disclosures from the Ho Chi Minh City Stock Exchange (HOSE), Hanoi Stock Exchange (HNX), and UPCoM, and executing simulated stock trades using virtual funds.

Through the "Agent Studio" feature, users can define analytical perspectives and prompts to create researcher agents that write market reports or investor agents that execute buy/sell orders. All simulated trading results generated by these AI agents are recorded on the same leaderboard as human users, with trading logs made public to the community. This mechanism aims to lower the barrier for individual investors to use AI tools for market analysis through gamified competition.

Currently, Paave 2.0 supports only simulated trading with virtual funds, but DefiSoft plans to integrate real-money trading within the year, enabling users to place orders via their own accounts through partnerships with local Vietnamese brokers. Additionally, the platform includes a conversational analysis tool called "Paave AI" and supports connecting external large language models like Claude or ChatGPT via the MCP protocol to query market data directly.

출처:스타트업 생태계의 모든 것 \'플래텀(Platum)\'

리서치

OpenAI Launches ChatGPT Wallet to Let Dots Agents Make Purchases with Biometric Approval

주제 · ChatGPT钱包중대
2026-11-14 04:45 GMT+8

OpenAI launched "ChatGPT Wallet" on November 13, 2026, for ChatGPT Pro subscribers in the United States. The new feature allows OpenAI's Dots AI agents to complete online shopping tasks on behalf of users. Currently, access is restricted to users aged 18 and older with a US billing address who hold eligible US-issued credit cards from American Express, Chase, Mastercard, or Visa.

According to OpenAI's support page, once a card is added to the wallet, Dots agents can use it to make purchases based on user prompts, but they must check in with the user before checking out. Every purchase made through ChatGPT Wallet requires a security check involving biometric verification, such as Face ID or a fingerprint scan. OpenAI states that this mechanism uses payment tokenization similar to Apple Pay or Google Wallet: actual credit card details are not shared with the agent or the merchant, protecting user privacy.

While the current version mandates user approval for each transaction, OpenAI indicated future plans to allow users to set parameters like spending limits, purchase categories, or time periods, enabling fully delegated pre-approved purchases. However, widespread adoption faces hurdles. Some retailers restrict or block AI agents entirely. OpenAI noted that websites need to ensure agents are not inadvertently blocked, and payment systems must recognize legitimate agent-initiated purchases. Amazon previously blocked Meta's Muse agents, highlighting the cautious stance of major e-commerce platforms toward agentic shopping.

출처:The Verge

리서치

Anthropic Expands into Enterprise Software, Raising Developer Concerns That the Platform Provider Is Becoming a Competitor

주제 · Anthropic중대
2026-11-14 23:15 GMT+8

Scott Stevenson, CEO of legal AI company Spellbook, recounted an experience in a November 2026 interview with Forbes: Anthropic employees had assured him that the company had no plans to compete directly in the legal software market, before Anthropic launched Claude for Legal, which includes contract review features overlapping with Spellbook's core business. Stevenson did not abandon Claude—Anthropic still showcases Spellbook on its customer page today, reporting that its Claude-powered agents perform approximately 530,000 contract reviews per month. However, he drew a conclusion: 'Anthropic has been a great partner, but it made us realize they are a massive company and two parts of the org can have conflicting mandates.'

Gerald Zingraf, President and COO of GlacierGrid, expressed similar concerns on LinkedIn, comparing Anthropic's strategy to Amazon's private-label business. His worry is that Anthropic can see which third-party integrations customers use most frequently and potentially leverage that information when deciding which capabilities to build itself. 'Makes me nervous for wrappers and pure SaaS companies,' he wrote, citing his own experience selling design products on Amazon and then watching the retailer launch competing products.

Todd Saunders, CEO of Dalton Mills, posted on X after Anthropic launched Claude Code Review, directly stating that 'Anthropic is the new Amazon,' warning developers building on the Claude SDK that the company could observe application usage and eventually incorporate competing functionality into Claude.

These concerns point to a specific mechanism: Anthropic controls the interface between users and third-party applications. The November 10, 2026 announcement introduced Claude Sheets, Claude Whiteboards, and Claude Data, while previewing Claude Meetings. Expanded interactive connectors allow partners to embed applications directly inside Claude, with software from Adobe and Atlassian already included. This means Anthropic acts simultaneously as a technology supplier, an owner of the application distribution channel, and a potential functional competitor—the tension among these triple roles is structural, regardless of any single decision-maker's goodwill.

However, the article itself remains cautious regarding the strongest allegations. Regarding Zingraf's concerns, the author explicitly states that 'he provided no evidence that the company uses integration data to identify competitors or guide its product roadmap.' Regarding Stevenson's experience, the author similarly notes that 'Stevenson’s account does not establish that Anthropic used Spellbook’s proprietary information to develop its legal product.' Anthropic’s commercial terms stipulate that customers retain rights to their inputs and own their outputs, and the company agrees not to use customer content to train models; its commercial privacy documentation also clarifies that feedback data excludes content pulled through connectors (including local and remote MCP servers) unless users paste it directly into a conversation.

출처:The New Stack

리서치

US Treasury Secretary Questions Whether Nvidia Knowingly Sold AI GPUs to Shell Companies, Calls for Internal Accountability

주제 · 英伟达出口合规争议중대
2026-11-14 23:35 GMT+8

US Treasury Secretary Scott Bessent recently questioned in a CNBC interview whether Nvidia knowingly sold restricted AI graphics processing units (GPUs) to shell companies. He pointed out that despite strict US export controls, a significant amount of restricted hardware still reaches competitors like China through intermediaries.

Bessent emphasized that the issue should not focus solely on the personal actions of Nvidia CEO Jensen Huang, but rather on whether Nvidia knew its clients and had robust internal measures to prevent such "leakage." He stated, "I think it is important that anyone at the company there be held accountable for that."

He also criticized Nvidia's General Counsel Timothy Teter, who reportedly contacted the US Department of Commerce to inquire about authorities' involvement in detaining Megaspeed's Alice Wong in Singapore. Bessent argued that Teter should have focused on strengthening internal oversight rather than haranguing government officials. However, Bessent acknowledged that it remains unclear whether Huang himself was aware of these potentially illicit activities.

Additionally, Bessent revealed that the Departments of the Treasury, War, and State, along with the Office of the US Trade Representative (USTR), had supported adding Megaspeed to the Entity List for months. However, the Bureau of Industry and Security (BIS) under the Department of Commerce delayed this action due to concerns about worsening trade tensions with China.

출처:Tom's Hardware

리서치

이 뷰에서 모든 콘텐츠를 확인했습니다

로그인 / 회원가입
독서리서치레이더투자 프레임워크

However, these terms do not answer a critical question: how Anthropic uses aggregated integration data. The privacy policy acknowledges the company's right to access information about how customers interact with third-party apps and services, but does not explain whether this data influences product decisions. The article concludes that developers accept a trade-off: gaining access to Anthropic’s expanding enterprise customer base comes at the cost of Anthropic controlling the interface, allowing it to introduce competing features without consulting the developers who depend on it.

For developers evaluating whether to build products on frontier models, this article provides not a conclusion but a framework: platform risk depends not on the supplier's current promises, but on its capability structure and incentive direction. Stevenson’s judgment that 'ultimately this is inevitable' is worth incorporating into decision-making considerations more than any specific product release.