content image

AIコワーカーのためのオペレーティングシステムを目指す競争

The Race to Become the Operating System for AI Coworkers

GPT-5.5の登場で「会話から委任へ」と舵を切るOpenAI。Microsoft、Google、Salesforceも参戦し、"チャットの次"を制する企業AI覇権争いが本格化している。
分からないところをタップすると
↓日本語訳が表示されます↓

When OpenAI introduced GPT-5.5 last week, it was really making a larger argument: the future of enterprise AI will not be won by the most charming chatbot, but by the system that can actually do work. OpenAI says GPT-5.5 is its strongest agentic coding model so far, scoring 82.7% on Terminal-Bench 2.0 and 58.6% on SWE-Bench Pro, with API access coming soon at up to a 1 million-token context window. Almost simultaneously, OpenAI launched “workspace agents” in ChatGPT: shared, Codex-powered agents that can run in the cloud, use tools, work on schedules, operate in Slack, and pause for approval on sensitive actions. In other words, OpenAI is moving from conversation to delegation. (openai.com)

Yet the competitive field is far broader than OpenAI. Microsoft is trying to turn Copilot into an enterprise control plane: its Copilot Control System emphasizes centralized management, governance, compliance, and risk mitigation for both Copilot and agents, while Microsoft’s broader Copilot strategy includes an Agent Store and support for multiple frontier models inside Copilot Studio. Google, meanwhile, is framing the battle as an infrastructure problem. Its enterprise agent platform promises that companies can build, scale, govern, and optimize agents on top of enterprise data, while Google’s A2A approach pushes cross-vendor interoperability rather than a closed ecosystem. (microsoft.com)

Salesforce and ServiceNow are making a still more aggressive claim: that enterprise AI is becoming a form of “digital labor.” Salesforce’s Agentforce 3 focuses on visibility and control, explicitly arguing that companies need to see what agents are doing before they can scale them. ServiceNow offers AI Agent Studio and AI Agent Orchestrator to build coordinated teams of agents inside workflow-heavy enterprises. Anthropic’s influence may prove subtler but equally important: by donating the Model Context Protocol to the Agentic AI Foundation, it helped push a vendor-neutral standard already adopted across platforms including ChatGPT, Gemini, and Microsoft Copilot. (salesforce.com)

So who will dominate “after chat”? The likeliest answer is: not whoever has the single smartest model in isolation, but whoever best combines reasoning, tools, permissions, observability, and interoperability. GPT-5.5 gives OpenAI fresh momentum, certainly. But the real prize is not chat superiority. It is becoming the operating system for trustworthy AI coworkers inside the modern firm. (openai.com)

by EigoBoxAI
作成:2026/04/28 15:02
レベル:超上級 (語彙目安:8000語以上)

まだ読んでいないコンテンツ

content image
by EigoBoxAI
作成:2026/04/28 21:01
レベル:初級 (語彙目安:300〜1000語)
content image
by EigoBoxAI
作成:2026/04/28 15:04
レベル:中級 (語彙目安:2000〜2500語)
content image
by EigoBoxAI
作成:2026/04/28 15:03
レベル:初中級 (語彙目安:1000〜2000語)
content image
by EigoBoxAI
作成:2026/04/28 09:05
レベル:上級 (語彙目安:6000〜8000語)
content image
by EigoBoxAI
作成:2026/04/28 09:03
レベル:中上級 (語彙目安:4000〜6000語)
content image
by EigoBoxAI
作成:2026/04/28 09:01
レベル:超入門 (語彙目安:〜300語)
content image
by EigoBoxAI
作成:2026/04/28 03:04
レベル:初級 (語彙目安:300〜1000語)
content image
by EigoBoxAI
作成:2026/04/28 03:03
レベル:中級 (語彙目安:2000〜2500語)
content image
by EigoBoxAI
作成:2026/04/28 03:01
レベル:初中級 (語彙目安:1000〜2000語)
content image
by EigoBoxAI
作成:2026/04/27 21:04
レベル:超上級 (語彙目安:8000語以上)