What is Responsible AI?

Responsible AI is a way of designing, deploying and monitoring AI systems so they are fair, transparent, accountable and safe.

What does AI governance involve for a small team?

A record of which AI tools you use, what data goes into them, who signs off on what, and what happens when something is wrong.

Independent Research on
Responsible AI

We study how AI is built, governed and used, and publish what we find: reviews of new research, a tracker of AI laws and standards, and commentary on AI in daily life.

Read our research

AI regulations tracker

AI news and policy updates

See all

Google Unveils Gemini 4 Argon After Delays

Alphabet's Google announced Gemini 4 Argon on September 30, its first frontier AI model in more than seven months, as the company seeks to close a gap with OpenAI and Anthropic (Reuters, 2026). The model is larger than Google's previous Pro line and is designed for complex workloads in software engineering, legal and financial work, and cybersecurity defense. A company spokesperson said Argon is comparable to OpenAI's Astra and Anthropic's Opus on key coding and cyber benchma

17 hours ago1 min read

OpenAI Launches Always-On Dots Agents

OpenAI unveiled always-on AI agents called dots on September 29 at its annual Developer Day in San Francisco, deepening its push into the enterprise market and setting up a direct challenge to Meta in the race to sell autonomous AI that can work with minimal supervision (Reuters, 2026). The agents, powered by GPT-6 Astra, operate across Slack and Teams and can draw on OpenAI's Codex and ChatGPT Work tools to research, analyze data, prepare documents, and build software. They

3 days ago2 min read

Microsoft Adds Coding and Agent Tools to Copilot

Microsoft unveiled several new capabilities for its Copilot app on September 25, including a coding tool and an always-on AI agent, as the company works to transform its assistant into a comprehensive workspace for office employees. The new Code tool lets users build apps, dashboards, and other software using natural language prompts. It runs on the same technology as Microsoft's GitHub Copilot and will reach early-access customers by the end of September, with broader previe

5 days ago1 min read

Amazon Bets $3 Billion on India Quick Commerce

Amazon will invest $3 billion to expand its India quick commerce business by 2030, two sources said, its largest bet yet on a sector where rivals have raced ahead by popularising deliveries within minutes. The instant delivery market has reshaped how urban Indians shop for everything from milk and chocolates to electronics. Walmart's Flipkart and Amazon, which dominate India's e-commerce space, are late entrants. The $19 billion sector is set to more than double to $41 billio

Oct 261 min read

Anthropic, Accenture Commit $2 Billion to AI Evaluation

Anthropic said on September 18 it will partner with Accenture to independently evaluate its frontier AI models, with each company committing at least $1 billion over the next five years to build capacity for the work. Accenture shares rose 7 percent in extended trading following the announcement. Accenture's specialist AI business, Faculty, will lead the partnership, evaluating and red-teaming Anthropic's models through alignment assessments and safeguard testing. The investm

Oct 152 min read

What We Study

Our research, evaluations and public conversations focus on how AI affects decisions, rights and daily life.

AI Governance and regulation

How AI laws and standards work in practice.

Open the tracker

AI ethics and accountability

Review of new research on fairness, accountability and the unintended effects of AI.

Read the reviews

AI in everyday life

How is AI changing pricing, hiring, health information and privacy.

Read commentary

Independent evaluations

Evaluating AI systems for performance, fit, risk and compliace with recognised standards.

Work with us

RAI Mark

An independent mark showing an AI system has been evaluated against responsible AI principles.

Coming soon

Podcast

Conversations with AI builders, policymakers and researchers on responsible AI in practice.

Listen on spotify

Research and commentary library

Explore the library

Everything we have published, by topic.

Latest commentary

See all

Turning Boring Documents into Audio Shows

Sitting down to read a dense, 50-page PDF report for work or school can feel like an insurmountable challenge. In a world saturated with short-form digital media, our collective attention spans are constantly tested by dry text, technical manuals, and corporate whitepapers. Skimming through walls of complex jargon often leads to poor retention and cognitive fatigue. While we know the information is important, our brains naturally resist processing hours of static text after a

3 days ago2 min read

Why You May Never Do Your Taxes Again

Dreading the end of the year because of the complicated, mind-numbing process of tracking expenses and filing taxes is a universal experience. Many of us spend hours digging through shoeboxes of paper receipts, scrolling through endless bank statements, and second-guessing every single deduction. The constant fear of making a simple mathematical error that triggers a costly government audit adds unnecessary stress to the entire process. For decades, tax preparation has felt l

7 days ago2 min read

AI Generating Retro Games on the Fly

Playing classic 8-bit or cyberpunk arcade games brings a wave of pure nostalgia. We fondly remember spending hours dodging pixelated obstacles, collecting glowing tokens, and chasing high scores. However, after beating the same static levels for years, even the most beloved titles lose their charm. For generations, gamers wished for an infinite stream of fresh challenges within their favorite retro worlds. Today, that dream of endless replayability is becoming a reality throu