| Bloomberg IntelligenceBloomberg LP · New York |
- ASKB — AI chat interface built into the Bloomberg Terminal (Feb 2026 beta) that draws on several AI models simultaneously to answer queries
- BloombergGPT — Bloomberg's own AI language model trained specifically on financial data; drives sentiment analysis, entity recognition, news categorisation, and headline generation
- AI Document Insights — AI tool that lets users ask plain-language questions about company filings and earnings call transcripts (Apr 2025)
- AI Earnings Call Summaries — automatically generated summaries of earnings calls, built into standard Bloomberg Terminal workflows
|
✓ Published
"Responsible AI principles" govern all product development; a dedicated Head of Responsible AI sits in the CTO office. Cited explicitly in ASKB product materials. Research on AI retrieval accuracy and financial content risk has been published publicly.
|
Active
Publishes research on AI safety risks in financial applications. Bloomberg Intelligence monitors global AI regulation on behalf of clients.
|
Bloomberg Terminal AI page ↗
Responsible AI research ↗
ASKB launch ↗
|
| EIUEconomist Intelligence Unit · London |
- Viewpoint AI — EIU's own AI-powered search and summary tool covering its full content library, built on Amazon's cloud AI platform (launched late 2025)
- EIU Q&A Copilot — internal AI assistant for editorial staff, hosted on Amazon's cloud; cut the time to deploy new features from 6 months to just days
- RavenPack / Bigdata.com API — EIU forecast data now available through connectors that allow clients to plug EIU analysis directly into their own AI applications (Mar 2026)
|
~ Partial
Subscriber agreements ban using EIU content to train AI models or feed it into other AI tools. No separate policy on how EIU staff may use AI has been made public.
|
Active
The Economist Group publishes research reports on AI adoption for corporate clients; AI governance is a key focus area across its research and product offerings.
|
Viewpoint AI launch ↗
RavenPack MCP deal ↗
AWS re:Invent 2025 ↗
|
| FT / FT LongitudeFinancial Times · London |
- OpenAI content deal (Apr 2024) — FT content licensed to train OpenAI's models; ChatGPT can now show attributed summaries of FT articles to users
- FT AI Playground — an internal sandbox where FT staff develop and test AI prompts using real FT editorial content before any public use
- Custom data pipelines — FT's data journalism team uses machine learning to automatically identify people, organisations, and patterns across large datasets
- AI article summaries — AI-generated summaries reviewed and edited by journalists, clearly labelled as AI-assisted on FT.com
|
~ Partial
AI-generated content is disclosed to readers through clearly labelled summaries. The OpenAI content deal is a public commercial agreement. No separate policy on how FT staff may use AI has been made public.
|
Active
FT journalists actively participate in debates on AI ethics in journalism. The OpenAI content deal has set a precedent for the industry. FT's editorial leadership has spoken publicly about standards for AI attribution and sourcing.
|
WAN-IFRA: How FT uses AI ↗
FT–OpenAI deal ↗
|
| Economist ImpactThe Economist Group · London |
- AWS Bedrock — shared Economist Group AI infrastructure on Amazon's cloud; handles content summarisation, translation, and personalised recommendations
- AI recommendations engine — directs subscribers to content tailored to their interests
No tools specific to Economist Impact have been publicly documented; it uses shared Economist Group infrastructure.
|
— Not found
Economist Impact has no separate AI policy. It operates under the broader Economist Group editorial standards and Amazon's responsible AI guidelines.
|
Limited
Publishes sponsored research on AI adoption for corporate clients. AI is mainly a revenue and research theme for Economist Impact rather than a public policy priority.
|
AWS re:Invent 2025 ↗
|
| Semafor IntelligenceSemafor · New York |
- Proprietary AI analysis tool — parses the full content of Semafor's global summits; identifies distinct claims, topics, stances, and evidence from each speaker; ranks themes by how broadly they are held; links findings to individual speakers, sessions, transcripts, and video clips
- First edition (Apr 2026) — analyzed 4,900 distinct claims from 300+ speakers at the Semafor World Economy summit in Washington, DC; editorial team reviews AI output before publication
The AI tool is proprietary and built specifically for Semafor's convening content. No other staff AI tools have been publicly documented.
|
— Not found
No published internal AI use policy found for Semafor staff.
|
— Not found
No public AI policy advocacy work identified. Semafor Intelligence is an editorial product, not a policy research unit.
|
Semafor Intelligence launch ↗
|