AI AND COMMUNICATION
AI is already at work in your organisation. Are you confident it's being used safely?
If your board or funder asked how, could you answer?
One in three Australian workers use AI without their employer knowing, often because they're unsure what's allowed. (Joseph Gabriel Lagonsin, IT Brief, 2026) And 83% of office workers in large companies say leaders follow different AI rules from everyone else. (Sean Mitchell, IT Brief, 2026)
In business, government and community services alike, unchecked AI can mean wrong information, private details exposed and lost trust. AI tools can also give wrong or out-of-date answers when people ask about organisations like yours.
This site shows real cases, what AI does well, the standard I work to, and a free self-test to check what AI tools say about your organisation.
WHY THIS MATTERS NOW
A revolution in my lifetime
When I started my communication degree in 1998, the year Google was founded, I majored in journalism and studied digital technology and multimedia every year alongside it. I've been an early adopter of every wave since: the internet, search, mobile phones, then social media in its early days, when Facebook was still mostly for university students and graduates. Working first in journalism and then in communications, I've watched each wave profoundly and forever change how people find, share and trust information.
Until now, the internet was the biggest disrupter I'd seen across work, life and how we communicate. AI is much bigger: the biggest technology revolution of my lifetime.
In January 2026, the head of the International Monetary Fund said AI is “like a tsunami hitting the labour market”. The IMF expects it to affect almost 40% of jobs worldwide. That rises to about 60% in advanced economies, such as the United States, Japan, the United Kingdom, Australia, New Zealand and Singapore. In China and other emerging markets, it's about 40%. (Tristan Bove, Fortune; Kristalina Georgieva, IMF)
Note: “advanced economies” and “emerging markets” are the IMF’s economic groupings, based mainly on income per person. They don’t measure how advanced a country’s technology is. China, for example, is highly advanced in technology. It ranks second in the world for AI, after the United States. (Stanford HAI, Global AI Vibrancy Tool 2025)
“Like a tsunami hitting the labour market”
Kristalina Georgieva, head of the International Monetary Fund, speaking on AI at the World Economic Forum in Davos, Switzerland, January 2026
I acknowledge the Traditional Owners and Custodians of the lands across Australia and pay my respects to Elders past and present.
IN THE NEWS
When AI gets it wrong
Before I show you what AI does well, and where it came from, I want to flag the risks, so you go in with your eyes wide open. Here goes: AI makes mistakes. It can sound confident while it makes up facts, sources and quotes. Some of the brightest minds have learned that the hard way, and their stories have made the news.
The risk grows when people assume AI is always right. They don't bring their own thinking and writing to it, and they don't check what it produces before they press publish. They don't follow standards and good governance in how they use it. In most of these cases, AI output wasn't checked well enough to catch the errors. In some, the people relying on the work weren't told AI was used.
SEPTEMBER 2026 · TASMANIA
A parole condition built on made-up case law
Tasmania's parole board barred a woman convicted of murder from speaking to the media about her claim that she was wrongly convicted. The document behind that condition, prepared by the board's secretary, cited fake, AI-generated cases that didn't exist. The Supreme Court of Tasmania lifted the condition, finding it was imposed without procedural fairness. The state's Attorney-General asked the board to explain. (Ethan James, Australian Associated Press)
JUNE 2026 · KPMG
A report on AI pulled after false claims
KPMG removed a published report on AI after organisations named in it, including the UK's National Health Service, said its claims about their AI use were untrue or misleading. A research group said the errors came from AI. KPMG began an internal investigation. (Anthony Ha, TechCrunch)
MARCH 2026 · HEALTH NEW ZEALAND
Clinical notes written with free AI tools
Health New Zealand found staff in a mental health and addiction service using free AI tools, such as ChatGPT and Gemini, to write clinical notes. A memo warned staff it could lead to disciplinary action, citing data security, privacy and accountability. A staff union said heavy workloads were driving people to AI, and called for training and approved tools. (RNZ)
OCTOBER 2025 · DELOITTE AUSTRALIA
A $440,000 government report with fake references
A report for a federal department included references to research that didn't exist and an inaccurate quote from a Federal Court judgment. AI use wasn't disclosed until a corrected version was published. Deloitte agreed to a partial refund. (Rod McGuirk, Associated Press)
SEPTEMBER 2024 · VICTORIAN CHILD PROTECTION
Private case details entered into ChatGPT
A child protection worker used ChatGPT to help write a report for the Children's Court. It included inaccurate information that downplayed risks to a child, and sensitive personal details were sent to an overseas company. The state's Information Commissioner ordered the department to block public AI tools for staff. (Ry Crozier, iTnews)
NOVEMBER 2023 · SENATE INQUIRY
False claims about real firms put on the public record
A group of academics used an AI tool to help write a submission to a Senate inquiry into the consulting industry. It invented allegations against KPMG and Deloitte, which were published on the parliamentary record. The academics apologised. (Consultancy.com.au)
When AI is used against you
This is the hardest reality. Like every useful invention, AI can be used for harm as well as good. So criminal syndicates and individuals in every region are exploiting it. The human suffering is global and significant. The United Nations Office on Drugs and Crime calls AI "a powerful force multiplier for criminal activities". Europol warns that "the very DNA of organised crime is changing".
The human toll runs both ways. Behind many scams are compounds in Southeast Asia where people are trafficked and forced to scam others. The UN estimates at least 120,000 people in Myanmar and 100,000 in Cambodia have been trapped this way, many facing torture and abuse (UN Human Rights Office, 2023).
A video call from your boss. A phone call from your grandson. A trusted doctor recommending a cure. Any of them could be fake. Deepfakes are videos, images and voices made with AI to look and sound like real people, and scammers use them to steal money and trust. Scams are now a global industry, and AI is making them harder to spot.
GLOBAL
57%
of adults worldwide faced a scam in the past year, and nearly 1 in 4 lost money. (Global Anti-Scam Alliance and Feedzai, 2025)
UNITED STATES
US$20.9 billion
lost to online crime in 2025, up 26%. More than 22,000 complaints, worth almost US$893 million, were linked to AI. (Sinisa Markovic, Help Net Security, on FBI figures)
AUSTRALIA
A$2.18 billion
lost to scams in 2025. The ACCC says AI is making scams more sophisticated. (Daniel Croft, Cyber Daily)
UNITED KINGDOM
£1.3 billion
stolen by fraudsters in 2025, across more than 4 million cases. UK Finance says criminals using AI are driving much of the growth. (Financier Worldwide)
MALAYSIA
RM1.58 billion
lost to online scams in 2024, and RM1.12 billion in the first half of 2025 alone. Police found deepfakes of Prime Minister Anwar Ibrahim promoting fake investments. (Benjamin Lee and others, The Star; Bernama)
CHINA
¥217 billion
in scam money frozen by police in 2025 (about US$31 billion), and 3.6 billion scam calls blocked. Courts say AI voice-cloning scams targeting older people are "alarmingly frequent". (Yang Zekun, China Daily; Cao Yin, China Daily)
Figures are total scam and fraud losses, cases or money frozen in each place, not AI alone. Each uses its own currency and measures.
JULY 2025 · AUSTRALIA
Deepfake doctors selling unproven cures
Deepfake videos misused the identities of trusted Australian health experts, including Dr Norman Swan, to sell unproven supplements for diabetes and other conditions. The Australian Medical Association called on the federal government for enforceable rules on online health advertising. (Hospital + Healthcare)
WHAT THESE CASES SHARE
Technology alone won't stop this
Authorities are fighting back on a massive scale. In 2025, China's police blocked 3.6 billion scam calls and froze ¥217 billion in scam money (Yang Zekun, China Daily). Australia's corporate regulator, ASIC, took down more than 19,400 online scams in a year, warning that AI is making them harder to detect (Damian Chmiel, Finance Magnates). But every case here worked because people trusted what they saw and heard. Pausing to check who you're really dealing with, another way, is still one of the best protections. Would your organisation know what to do?
How I can help →I acknowledge the Traditional Owners and Custodians of the lands across Australia and pay my respects to Elders past and present.
THE OTHER SIDE
What AI does well
AI isn't new. John McCarthy, a professor at Dartmouth College in the United States, coined the term in 1955. The field began there a year later (IndiaAI). Most people have used AI for years without realising, in email spam filters and in medical equipment. What changed in late 2022 was generative AI: tools like ChatGPT that create text, images and sound by predicting what sounds right. That's what makes them so useful, and why they can be so confidently wrong at times.
Used with care, AI does real good. It can ease human suffering. Bring hope to hopeless situations.
These uses and their results are published.
MARCH 2025 · UNITED STATES
A second chance at life
AI tools built for drug and medical research can help find cures for diseases. They work best when skilled and caring professionals use them, keep questioning, and check the results. Together, they can open up new options for people living with chronic, rare and terminal conditions.
“How many other cures are hiding in plain sight?”
The question researchers are now asking (Kate Morgan, The New York Times)
Joseph Coates, 37, had a rare blood disorder called POEMS syndrome. He was too sick for the stem cell transplant he needed, and was told to choose where to spend his final days. But his girlfriend wasn't ready to give up. She contacted Dr David Fajgenbaum, a University of Pennsylvania doctor the couple had met at a rare disease summit. He proposed a drug combination no one had tried for Joseph's condition, suggested by an AI tool that matches about 4,000 existing medicines against 18,500 diseases. Within a week Joseph began to improve. Four months later he had his transplant, and he is now in remission.
Finding new uses for existing medicines isn't new, but AI is speeding it up. Scientists around the world now use it to find treatments for rare diseases, for people with few other options. (Kate Morgan, The New York Times)
JANUARY 2026 · SWEDEN
More breast cancers found at screening
In a trial with more than 100,000 women, AI-supported screening found more cancers than standard double reading by two radiologists, with no rise in false alarms. Earlier results from the same trial showed it cut screen-reading workload by 44%. (Jessie Gommers and others, The Lancet)
APRIL 2026 · UNITED STATES
Less time on notes for clinicians
Across more than 8,500 clinicians at five health systems, using an AI scribe was linked to 16 fewer minutes of documentation for every eight hours with patients, about 10% less. (Lisa Rotenstein and others, JAMA)
2024 · AUSTRALIAN PUBLIC SERVICE
Time back for higher-value work
In a six-month government trial of an AI assistant, staff estimated saving up to an hour on tasks such as summarising and first drafts. 64% of managers saw gains in efficiency and quality, although only about a third of staff used it daily. (Digital Transformation Agency)
SINCE 1995 · MEDICAL DEVICES
Three decades of AI in health care
The US Food and Drug Administration authorised its first AI-enabled medical device in 1995. By August 2024 it had authorised 950, most of them in radiology. (Elise Reuter and Jasmine Ye Han, Healthcare Dive)
SINCE 2017 · EMAIL
Keeping spam out of your inbox
Google says machine learning helps Gmail block spam and phishing emails with more than 99.9% accuracy. It's a kind of AI most people rely on every day without thinking about it. (Andy Wen, Google)
WHAT GOOD USE LOOKS LIKE
Check. Question. Act.
In these examples, the best results came when trained people used AI, measured the results, checked the work and acted on what they found. Judgement makes the difference.
How I use AI →I acknowledge the Traditional Owners and Custodians of the lands across Australia and pay my respects to Elders past and present.
MY STANDARD
Your thinking first. AI second.
I bring my own ideas and thinking to every piece of work. AI helps me work faster. These are the rules I follow.
01
My thinking comes first
I start with my own ideas, experience and judgement. AI comes in after that, never instead of it.
02
I check everything
AI makes mistakes. I edit, revise and fact-check all AI output, including sources, tone and wording, before anything is used.
03
Your information stays protected
I don't put client, personal or confidential information into public AI tools without your agreement.
04
Community data and stories are respected
I follow Indigenous data sovereignty principles. Communities decide how their information and stories are collected, shared and used.
05
A person signs off
Every piece of work is reviewed and approved by me, and by you where it's your content.
06
I'm open about it
If you'd like to know whether AI was used on your work, I'll tell you.
This website and its self-test were designed and built by me, with AI's help. Last updated: 4 November 2026
HOW I CAN HELP
People checking the work
Every case on this page comes back to the same thing: people checking the work, protecting private information and being open about how it was done. That's the work I help organisations with.
YOUR TEAM
Using AI well at work
Practical sessions for your staff on checking AI output, protecting private information and being open about when AI is used. Built around your own work, not generic examples.
YOUR CONTENT
A second set of eyes
An independent check of important reports, statements and public content before they go out, for accuracy, sources, tone and plain language.
YOUR PUBLIC PICTURE
What AI says about you
An AI Visibility Check: a fuller review of what AI tools tell people about your organisation, what's wrong or missing, and what to fix first. The free self-test is a good place to start.
WHEN IT GOES WRONG
Ready before it happens
Readiness for the incidents boards now worry about, from a data breach to a deepfake of your CEO: who decides, who needs to know, and in what order in the first 48 hours. Practised before it's real.
We'll talk about where you are now and what would help most. No obligation.
I acknowledge the Traditional Owners and Custodians of the lands across Australia and pay my respects to Elders past and present.
FREE SELF-TEST
What Does AI Say About You?
More people now ask AI tools about services before they call, visit or fund them. This short check helps you see whether those tools are giving people a clear, current and fair picture of your organisation.
It takes about 15 minutes. Your answers stay in your browser. Nothing is sent or stored.
Step 1: Ask the AI tools
Open two or three AI tools, such as ChatGPT, Gemini, Copilot or Perplexity. Ask each one these questions, using your organisation's name.
Step 2: Answer 12 questions
Choose the answer closest to what you found.
YOUR RESULTS
Get a free results guide that explains what each area means for your organisation, plus a Get Found by AI starter checklist.
Get my guide and checklistThis is a quick self-check, not a full review. AI tools change their answers often, so it's worth checking again every few months.
'+n+'. '+q+'
RESULTS GUIDE
Get your guide and checklist
Leave your details to open your free results guide and Get Found by AI starter checklist straight away. The guide explains what each area of the self-test means for your organisation, and the checklist shows where to start. I may follow up once to offer a free 30-minute conversation.
Your details are used only for this. See my Privacy Policy.
I acknowledge the Traditional Owners and Custodians of the lands across Australia and pay my respects to Elders past and present.
YOUR RESULTS GUIDE AND CHECKLIST
What Does AI Say About You?
Thank you for taking the self-test. This guide explains what your score means and why each of the four areas matters. Your Get Found by AI starter checklist follows.
Your score
Your score out of 24 places you in one of three bands.
| Score | Result | What it suggests |
|---|---|---|
| 19–24 | Strong foundations | AI tools seem to give people a fairly clear and accurate picture of you |
| 11–18 | Some gaps to look at | People may be getting an incomplete or out-of-date picture of you |
| 0–10 | Worth acting on | People may be getting wrong, old or missing information about you |
Your lowest-scoring area is usually the best place to start.
1. FindabilitySTART HERE
What it means: whether AI tools know your organisation exists, what you do and where to find you.
Why it matters: many people now ask an AI tool before they search, call or visit. If the tool doesn't know you, it may point them to another service, or give no answer at all. For a business, that can mean lost customers. For health and community services, it can mean someone doesn't get help they need, or turns up in the wrong place.
Signs worth noticing:
- The tools couldn't find you, or mixed you up with another organisation with a similar name
- Your main services were missing or described vaguely
- Your address, phone number or opening hours were wrong or missing
2. AccuracySTART HERE
What it means: whether what AI tools say about you is correct, current and fair.
Why it matters: AI tools can repeat old information, mix up details, or fill gaps with things that sound right but aren't. People often trust these answers without checking. Wrong or outdated information can confuse customers and clients, worry funders and take time to correct.
Signs worth noticing:
- Programs you've closed, staff who've left or places you no longer work were still mentioned
- Details you didn't recognise, or that seemed made up
- A tone or description you wouldn't want a customer, community member, funder or journalist to read
3. Your own informationSTART HERE
What it means: how clear, current and easy to find your own website and listings are.
Why it matters: AI tools build their answers from information already online. When your own information is unclear, old or scattered, the tools rely more on other sources, which you don't control. Clear information from you gives people, and the tools they use, a better starting point.
Signs worth noticing:
- Your website doesn't say plainly who you are, who you serve and what you do
- Your Google listing or directory listings haven't been checked for a while
- The AI tools named other websites as their sources, rather than yours
4. ReadinessSTART HERE
What it means: whether your organisation has people and rules in place for how it shows up online and how staff use AI tools.
Why it matters: without someone looking after this, problems can sit unnoticed until a client, customer, funder or journalist raises them. Without clear guidance, staff may put confidential information into public AI tools.
Community data and stories: for Aboriginal community controlled organisations and many community services, this also means respecting Indigenous data sovereignty. Communities have the right to decide how their information and stories are collected, shared and used, including by AI tools.
Signs worth noticing:
- No one is clearly responsible for checking what's said about you online
- Staff have no guidance on what must never go into AI tools
- There are no agreed rules for sharing your community's information and stories
STARTER CHECKLIST
Get Found by AI
Start with your lowest-scoring area. Where a step needs your web provider, it gives you the question to ask.
1. Check your website isn't blocking AI search
Some website security services now block AI tools by default, so your site may be invisible to AI search without anyone deciding that.
- Ask your web provider: "Is our website blocking AI search tools, such as ChatGPT search, Perplexity or Claude?"
- Decide what you want: you can let AI search tools find your site, while still saying no to your content being used to train AI.
- Community controlled organisations: make this a board or community decision, in line with your data sovereignty principles.
- Put your decision in writing in your website terms.
2. Make sure search engines have your details right
AI tools often draw on search engines. Microsoft's Copilot, for example, uses the same foundation as Bing search.
- Claim and check your Google Business Profile: name, address, phone, hours and main services
- Do the same in Bing Places
- Add each site or clinic you run, including outreach and visiting services
- Ask your web provider: "Have we registered our site with Google Search Console and Bing Webmaster Tools?"
- Ask your web provider: "Does our website tell search engines our name, address and contact details in the background?" This is called structured data.
3. Say clearly who you are on your own website
- Your About page says, in plain words, who you are, who you serve, where, and what you offer
- Each main product or service has its own short description, written for the people who use it
- Regional and remote services list the communities or regions they cover
- Community controlled organisations say so, and include other details only where the community agrees
- Old programs, past staff names and closed sites are removed or updated
- Contact details match your Google and Bing listings exactly
4. Keep your name and details the same everywhere
- List every place you appear online, such as directories, member listings, review sites and funder listings
- Use the same name, and the same one-line description, everywhere
- Link back to your own website from every listing
- Fix or ask to remove any listing that is old or wrong
5. Look after it over time
- Give one person responsibility for checking what's said about you online
- Every three months, ask two or three AI tools the four self-test questions and save the answers with the date
- When something changes, update your website and listings in the same week
- Give staff simple guidance on what must never go into AI tools: client, customer, health, personal and commercially confidential information
- Publish short, plain answers to the questions people ask you most
What to do next
- Keep a record. Save what each AI tool said about you, with the date, so you can compare next time.
- Share it. Talk through your results with your leadership team, so someone owns this area.
Free 30-minute conversation. If you'd like to talk through your results and work out your next steps, I offer a free conversation with no obligation.
AI Visibility Check. For a deeper look, I review what several AI tools say about your organisation, compare it with your own information, and give you a written summary with priorities for action.
Get in touchAbout Vienna Senima Richards
I'm a communications, media and engagement professional with more than two decades of experience across government, public health, broadcasting and community organisations in Australia, New Zealand and the Pacific. I help organisations explain complex information clearly and build trust with the people they serve.
This guide and checklist are general information to help you reflect on your self-test results. They are not a full review, or technical or legal advice for your organisation.
Back to the self-test · viennarichards.com
© 2026 Vienna Senima Richards. All Rights Reserved.
