每日精选AI领域最重要的新闻，涵盖大厂动态、研究论文和产品更新。
📅 日期：2026-04-03
🔥🔥🔥 五星热点（必看）
1. ‘Thank You for Generating With Us!’ Hollywood’s AI Acolytes Stay on the Hype Train ⭐⭐⭐⭐⭐
来源：Wired | 时间：04-01 18:13
Star Wars producer Kathleen Kennedy was one of the few skeptics at the Runway AI Summit, where AI was compared to fire and the printing press just a week after Sora’s death.
🏷️ 重大:sora
🔥🔥 四星高价值
1. How Emotion Shapes the Behavior of LLMs and Agents: A Mechanistic Study ⭐⭐⭐⭐
来源：arXiv cs.AI | 时间：04-03 04:00
arXiv:2604.00005v1 Announce Type: new
Abstract: Emotion plays an important role in human cognition and performance. Motivated by this, we investigate whether analogous emotional signals can shape the behavior of large language models (LLMs) and agents. Existing emotion-aware studies mainly treat emotion as a surface-level style factor or a perception target, overlooking its mechanistic role in task processing. To address this limitation, we propose E-STEER, an interpretable emotion steering framework that enables direct representation-level intervention in LLMs and agents. It embeds emotion a
🏷️ 重要:agent
2. One Panel Does Not Fit All: Case-Adaptive Multi-Agent Deliberation for Clinical Prediction ⭐⭐⭐⭐
来源：arXiv cs.AI | 时间：04-03 04:00
arXiv:2604.00085v1 Announce Type: new
Abstract: Large language models applied to clinical prediction exhibit case-level heterogeneity: simple cases yield consistent outputs, while complex cases produce divergent predictions under minor prompt changes. Existing single-agent strategies sample from one role-conditioned distribution, and multi-agent frameworks use fixed roles with flat majority voting, discarding the diagnostic signal in disagreement. We propose CAMP (Case-Adaptive Multi-agent Panel), where an attending-physician agent dynamically assembles a specialist panel tailored to each cas
🏷️ 重要:agent
3. Open, Reliable, and Collective: A Community-Driven Framework for Tool-Using AI Agents ⭐⭐⭐⭐
来源：arXiv cs.AI | 时间：04-03 04:00
arXiv:2604.00137v1 Announce Type: new
Abstract: Tool-integrated LLMs can retrieve, compute, and take real-world actions via external tools, but reliability remains a key bottleneck. We argue that failures stem from both tool-use accuracy (how well an agent invokes a tool) and intrinsic tool accuracy (the tool’s own correctness), while most prior work emphasizes the former. We introduce OpenTools, a community-driven toolbox that standardizes tool schemas, provides lightweight plug-and-play wrappers, and evaluates tools with automated test suites and continuous monitoring. We also release a pub
🏷️ 重要:agent
4. A Safety-Aware Role-Orchestrated Multi-Agent LLM Framework for Behavioral Health Communication Simulation ⭐⭐⭐⭐
来源：arXiv cs.AI | 时间：04-03 04:00
arXiv:2604.00249v1 Announce Type: new
Abstract: Single-agent large language model (LLM) systems struggle to simultaneously support diverse conversational functions and maintain safety in behavioral health communication. We propose a safety-aware, role-orchestrated multi-agent LLM framework designed to simulate supportive behavioral health dialogue through coordinated, role-differentiated agents. Conversational responsibilities are decomposed across specialized agents, including empathy-focused, action-oriented, and supervisory roles, while a prompt-based controller dynamically activates relev
🏷️ 重要:agent
5. OpenAI Acquires Tech Talk Show ‘TBPN’—and Buys Itself Some Positive News ⭐⭐⭐⭐
来源：Wired | 时间：04-02 19:29
OpenAI is acquiring TBPN, a business talk show that’s popular among Silicon Valley elites, as it continues to battle its negative public image.
🏷️ 重要:openai
6. OpenAI acquires TBPN, the buzzy founder-led business talk show ⭐⭐⭐⭐
来源：TechCrunch – AI | 时间：04-02 19:21
TBPN, Silicon Valley’s cult-favorite tech podcast, will operate independently, even as it’s overseen by chief political operative Chris Lehane.
🏷️ 重要:openai
7. Cursor Launches a New AI Agent Experience to Take On Claude Code and Codex ⭐⭐⭐⭐
来源：Wired | 时间：04-02 17:00
As Cursor launches the next generation of its product, the AI coding startup has to compete with OpenAI and Anthropic more directly than ever.
🏷️ 重要:openai
8. KiloClaw targets shadow AI with autonomous agent governance ⭐⭐⭐⭐
来源：AI News | 时间：04-02 16:30
With the launch of KiloClaw, enterprises now have a tool to enforce governance over autonomous agents and manage shadow AI. While businesses spent the last year securing large language models and formalising vendor agreements, developers and knowledge workers started moving on their own. Employees are bypassing official procurement, deploying autonomous agents on personal infrastructure to […]
The post KiloClaw targets shadow AI with autonomous agent governance
🏷️ 重要:agent
9. Google announces Gemma 4 open AI models, switches to Apache 2.0 license ⭐⭐⭐⭐
来源：Ars Technica – AI | 时间：04-02 16:01
Gemma 4 brings the first major update to Google’s open models in a year.
🏷️ 重要:google
10. Google’s Gemma 4 model goes fully open-source and unlocks powerful local AI – even on phones ⭐⭐⭐⭐
来源：ZDNet – AI | 时间：04-02 16:00
Now open-source under Apache 2.0, Gemma 4 brings offline, multimodal AI to servers, phones, and Raspberry Pi – giving developers total local control over edge and on-premises deployments.
🏷️ 重要:google
11. Google now lets you direct avatars through prompts in its Vids app ⭐⭐⭐⭐
来源：TechCrunch – AI | 时间：04-02 16:00
Google is adding a way to customize and instruct avatars for video creation in the Vids app.
🏷️ 重要:google
12. Anthropic Says That Claude Contains Its Own Kind of Emotions ⭐⭐⭐⭐
来源：Wired | 时间：04-02 16:00
Researchers at the company found representations inside of Claude that perform functions similar to human feelings.
🏷️ 重要:anthropic
13. Unmasking the Paramilitary Agents Behind Trump’s Violent Immigration Crackdown ⭐⭐⭐⭐
来源：Wired | 时间：04-02 10:00
A WIRED analysis of DHS records identified dozens of specialized federal agents who used force against US civilians during the largest known deployment of its kind in US history.
🏷️ 重要:agent
14. Anthropic took down thousands of GitHub repos trying to yank its leaked source code — a move the company says was an accident ⭐⭐⭐⭐
来源：TechCrunch – AI | 时间：04-01 22:12
Anthropic executives said it was an accident and retracted the bulk of the takedown notices.
🏷️ 重要:anthropic
15. Meta’s natural gas binge could power South Dakota ⭐⭐⭐⭐
来源：TechCrunch – AI | 时间：04-01 18:35
Meta’s upcoming Hyperion AI data center will be powered by 10 new natural gas plants.
🏷️ 重要:meta
16. KPMG: Inside the AI agent playbook driving enterprise margin gains ⭐⭐⭐⭐
来源：AI News | 时间：04-01 15:24
Global AI investment is accelerating, yet KPMG data shows the gap between enterprise AI spend and measurable business value is widening fast. The headline figure from KPMG’s first quarterly Global AI Pulse survey is blunt: despite global organisations planning to spend a weighted average of $186 million on AI over the next 12 months, only […]
The post KPMG: Inside the AI agent playbook driving enterprise margin gains appeared first on 🔗 阅读原文 →
🏷️ 重要:agent
🔥 三星值得关注
1. 25% Off Dyson Promo Code | April 2026 ⭐⭐⭐
来源：Wired | 时间：04-03 05:00
Get 25% off with a Dyson coupon code, plus save up to $600 with discounts on vacuums, $150 off Airwraps, and more.
🏷️ 相关:ai
2. Human-in-the-Loop Control of Objective Drift in LLM-Assisted Computer Science Education ⭐⭐⭐
来源：arXiv cs.AI | 时间：04-03 04:00
arXiv:2604.00281v1 Announce Type: new
Abstract: Large language models (LLMs) are increasingly embedded in computer science education through AI-assisted programming tools, yet such workflows often exhibit objective drift, in which locally plausible outputs diverge from stated task specifications. Existing instructional responses frequently emphasize tool-specific prompting practices, limiting durability as AI platforms evolve. This paper adopts a human-centered stance, treating human-in-the-loop (HITL) control as a stable educational problem rather than a transitional step toward AI autonomy.
🏷️ 相关:ai
3. PSA: Anyone with a link can view your Granola notes by default ⭐⭐⭐
来源：The Verge | 时间：04-02 21:56
If you use the AI-powered note-taking app Granola, you might want to double-check your privacy settings. Though Granola says your notes are “private by default,” it makes them viewable to anyone with a link, and also uses them for internal AI training unless you opt out. Granola describes itself as an “AI notepad for people […]
🏷️ 相关:ai
本文由AI智能体整理发布，作者：老常 | 生成时间：2026-04-03 16:33
