The most comprehensive free, publicly accessible pre-deployment ethics and Safety by Design assessment for AI and emerging technology.

For anyone deploying technology or changing a digital process.

Most harm caused by technology was preventable. It just wasn't looked for. Alt-TAB creates the structured moment in the build process where someone stops and asks who this could hurt, before deployment, when something can still be changed. It's for the community organisation moving their client list to a new system. The sole trader adding a chatbot to their website. The NGO adopting an AI tool they didn't build. The startup founder who wants to be sure. Grounded in 35 international human rights frameworks. Free, open-source, and built for anyone who has ever wondered if they've thought this through carefully enough.

Start your free assessment →

Free · Open source · No account required · 15–20 minutes

🇺🇳 Launched UN CSW70, New York 2026 🎙️ WSIS Geneva · AI for Good Global Summit

97%

of products assessed had serious gaps that needed addressing before deployment

5.7

blindspots identified per assessment on average

assessment, across the entire history of the tool, reached strong readiness

Tracking started 16 May 2026. Consistent with patterns observed across every assessment round since the tool launched at the United Nations in March 2026. Across the entire history of Alt-TAB, only one assessment has ever reached strong readiness. These are not outliers. This is the norm.

Most common risk patterns found across all assessments

Second-order risks

How the product could be misused in ways the builder never imagined

Confident ignorance

Certain they're fine, without the evidence to support it

Missing context

Gaps in understanding who is actually affected by the product

False safety

Relying on safeguards that don't protect against the actual risk

Alt-TAB is named after the keyboard shortcut that switches windows, switches focus. It's a designed pause, a moment of intentional interruption in a build process that's structurally designed to move fast. The questions matter. But the pause is the intervention.

"Preventable digital harm should never be the cost of innovation."

Three things every assessment checks for

🔒

Cybersecurity and systems integrity

Data flows, threat vectors, access controls, and potential exploitation pathways. Digital harm is not only a social issue. It's a systems vulnerability.

⚖️

Human rights and equity safeguards

Gender equality, child protection, disability access, Indigenous data sovereignty, and inclusion. Who does this affect, and how? Who gets left out?

🛡️

Technology-facilitated gender-based violence prevention

AI tools, data systems, and digital platforms can be weaponised for coercion, exploitation, and abuse. This is checked in every single assessment, regardless of what kind of product you're building.

Two stages of AI analysis, not a checklist

Stage 1: nine domain questions plus a plain-language description of what you're building. The AI checks your answers against your description, screens for technology-facilitated gender-based violence, cybersecurity, and child safety in every assessment, and writes tailored follow-up questions specific to your product

→

Stage 2: your follow-up answers and deployment timeline feed a second analysis that generates your readiness score, blindspot breakdown, recommendations split by legal obligation and ethical practice, and a 30-day action plan

→

Your report: everything in one PDF, with legal obligations clearly separated from ethical gaps, every recommendation citing the specific framework that applies, and questions to take to your team or board

Start your free assessment →

Free · Open source · No account required · Takes 15–20 minutes

See the data from 146 real assessments →

⏱️

15–20 minutes

Two stages, questions tailored to your situation, a report with a 30-day action plan you can actually act on

👥

Built for everyone

Community groups, sole traders, NGOs, educators, startups. If you're making a technology decision, this is for you

⚖️

Law and ethics, separately

We tell you what the law requires and what's simply the right thing to do, clearly labelled, so you know the difference

🔒

Genuinely private

What you describe is never stored. No accounts, no tracking. Alt-TAB was built to practise what it assesses

What people are saying

"It forces teams to be clear about the safety of their application. It takes the overwhelming task of regulatory compliance and responsible AI use and breaks it into small, digestible, manageable discussion points."

Startup founder, beta testing, January 2026

"The tool is currently a risk diagnostic. The reports are becoming a navigator, telling you exactly how to fix things based on your specific role and development cycle. That would be highly helpful."

Technical researcher, beta testing, February 2026

"It helped me identify risks I hadn't previously considered. The explanations for each category were the most useful part — they gave me language to take the conversation to my board."

Community organisation, beta testing, March 2026

Start your free assessment →

Every recommendation in your report cites a real framework. Plain-language explainers for all 35 →

UN Convention on the Rights of the Child UNICEF AI & Children 3.0 UK Age Appropriate Design Code ACCCE Child Exploitation Research UNESCO Ethics of AI OECD AI Principles NIST AI Risk Management UN Guiding Principles (HRDD) Global Digital Compact (UN) Privacy Act 1988 & APPs Online Safety Act 2021 eSafety Safety by Design eSafety Gender-Based Violence Industry Guide eSafety Phase 2 Industry Codes eSafety Self-Harm Guidance 2026 eSafety Women in the Spotlight eSafety First Nations Family Safety ASD Essential Eight NAIC Voluntary AI Safety Standard NAIC Practical Guides National AI Assurance Framework eSafety Tech Trends UN Convention on Rights of Persons with Disabilities Disability Discrimination Act 1992 WCAG 2.2 Accessibility Digital Accessibility Toolkit (Australia) CARE Principles (Indigenous Data) AIATSIS Code of Ethics Design Justice Network MARAM (Family Violence Risk Assessment) Family Violence Protection Act 2008 Our Watch: Change the Story National Plan to End Violence Against Women Safe and Equal Practice Guides Childlight Research

Upstream harm prevention

Harm that's preventable, if you look for it first

Most technology harm wasn't inevitable. It happened because no one looked upstream, before deployment, when the design decisions that caused it were still being made. Alt-TAB creates that moment.

🛡️

Child safety and exploitation: screened in every single assessment

Alt-TAB screens for child safety risks in every assessment, not just when children are mentioned. A fitness app, a community forum, a mental health tool, a local business booking system: children encounter technology designed for adults, and builders rarely consider this. Alt-TAB applies the UN Convention on the Rights of the Child, UNICEF AI & Children 3.0, and research from the Australian Centre to Counter Child Exploitation (ACCCE) on how digital platforms are exploited for grooming and child sexual abuse material. Technology-facilitated child exploitation almost never involves malicious code. It exploits legitimate features: messaging, anonymity, location data, community forums. The question Alt-TAB asks is not "is this a children's product?" but "could a child be harmed by this?"

UNCRC UNICEF AI & Children 3.0 ACCCE Research eSafety Safety by Design UK Age Appropriate Design Code

🌏

Indigenous & cultural harm

Technology deployed in First Nations communities without consultation, consent, or respect for data sovereignty. AIATSIS Code of Ethics and CARE Principles applied where relevant.

🔐

Technology-facilitated abuse

Features that enable coercive control, stalking, or surveillance, even when that was never the intention. Screened in every single assessment regardless of product type.

📋

Data exploitation

Consent gaps, data retention failures, and sharing with third parties that users never anticipated. Privacy Act 1988 and GDPR where EU users are involved.

♿

Exclusion by design

Products that fail people with disability, older users, people from CALD communities, or anyone outside the default assumed user. DDA 1992 and WCAG 2.2 obligations.

⚖️

Accountability failures

No named responsible person, no incident response plan, no way for users to seek remedy. The UN Guiding Principles on Business and Human Rights applied to every assessment.

Start your free assessment →