Daily AI brief

AI News & Tools: Training-data scrutiny, infrastructure funding and enterprise controls

Court filings expose internal concerns about training data, Crusoe raises $3.9 billion, and new legal and Copilot capabilities put enterprise workflows in focus.

2026-09-18·Updated 2026-09-18·Primary sources shown for every item

01Policy & safety
Unsealed filings reveal Microsoft's internal objections to OpenAI's data practices
Newly unsealed court filings reveal Microsoft privately characterized OpenAI's data practices as theft, TechCrunch reports. The reporting also describes both companies scraping paywalled Times content, building datasets from it and internally warning about damage to publishers.
Why it mattersFor teams assessing training-data provenance, the reporting makes internal knowledge and objections relevant alongside the collection practices themselves. Those disclosures are evidence in a dispute, not a judicial determination.
Read analysis
SourcesTechCrunch (opens in a new tab)
02Business
Crusoe raises $3.9 billion for data centers and modular AI factories
Crusoe has raised $3.9 billion to build large data centers and small modular AI factories, according to TechCrunch. The round values the company at $30.9 billion.
Why it mattersCompute buyers need usable infrastructure, so the practical question is how this financing translates into facilities they can access. The two construction formats make deployment details relevant to capacity planning.
Read analysis
SourcesTechCrunch (opens in a new tab)
03Products
OpenAI introduces Astra for Law with firm workflows and connected legal data
OpenAI introduced Astra for Law, describing an offering with custom firm workflows and connected legal data sources. The company also promotes controls for confidential client work, but does not specify those controls or availability in its announcement.
Why it mattersLaw firms evaluating AI need to assess how client information moves through connected systems and which workflow steps they can configure. The announcement identifies those buying criteria without establishing how the controls operate.
Read analysis
SourcesOpenAI (opens in a new tab)
04Products
GitHub adds Copilot CLI customization activity to its usage metrics API
GitHub expanded Copilot CLI reporting in its usage metrics API to include agentic customization activity. The added coverage includes skills, custom agents, MCP servers, slash commands and plugins.
Why it mattersTeams maintaining shared agent configurations can use adoption information when deciding where to invest support and documentation. Usage counts alone do not establish whether a customization improves task outcomes.
Read analysis
Related directory entriesGitHub Copilot — The most widely deployed coding assistant, available in most major editors.
SourcesGitHub Changelog (opens in a new tab)
05Products
GitHub adds feature engagement to the Copilot impact dashboard
GitHub's Copilot impact dashboard now shows how many active users regularly use key features. GitHub says the view helps enterprise administrators identify widely adopted experiences and those that may need more attention.
Why it mattersAdministrators can use feature engagement to target onboarding and support. Interpreting the figures as productivity gains would require separate evidence about work completed and its quality.
Read analysis
Related directory entriesGitHub Copilot — The most widely deployed coding assistant, available in most major editors.
SourcesGitHub Changelog (opens in a new tab)

Unsealed filings reveal Microsoft's internal objections to OpenAI's data practices

Crusoe raises $3.9 billion for data centers and modular AI factories

OpenAI introduces Astra for Law with firm workflows and connected legal data

GitHub adds Copilot CLI customization activity to its usage metrics API

GitHub adds feature engagement to the Copilot impact dashboard