Claude's maker, Anthropic, has published the first wave of its Public Record survey, a large study of how people actually feel about AI. The numbers are worth a careful read for any Australian business owner planning an AI rollout. They show a public that is genuinely hopeful about what AI can do and genuinely wary of the companies building it. Anthropic calls this series the Public Record and intends to repeat it, which makes it a useful baseline to watch over time. That gap between hope and trust is the thing you have to manage when you bring AI into a team or put it in front of customers.
What the survey found
Anthropic fielded the first wave between November and December 2025, asking roughly 52,000 Americans about their hopes and fears for AI. A few results stand out and deserve attention from anyone introducing the technology to a workforce.
The top hope was curing diseases such as cancer or Alzheimer's, named by 48 percent, followed by helping people with disabilities at 36 percent.
The most common fear, in every single state, was job loss at 64 percent, ahead of cognitive dependency at 56 percent and misinformation at 52 percent.
More than 70 percent thought government should play a role in regulating AI, a view that held across party lines.
Only 15 percent said they trust AI companies to make decisions about how AI is developed and used.
The survey is American, so treat it as a signal rather than a local measurement. Even so, the pattern is consistent enough that it would be unwise to assume Australian staff and customers feel very differently about the same questions.
Why a US survey matters in Australia
Australian attitudes to AI track the same broad shape: real enthusiasm for the upside, real anxiety about jobs, and limited trust in vendors. The policy conversation in Canberra mirrors the survey's finding that people want accountability, with privacy, child safety, and liability for harm near the front of mind. Australian buyers also sit under their own rules, including the Privacy Act and various sector codes, so a careful rollout is not optional for many firms. For a business, the practical reading is simple. The people you most need on board with an AI project, your own staff and your customers, are starting from a position of caution.
The trust gap is a commercial problem
It is tempting to file public sentiment under ethics and move on. That is a mistake. When only 15 percent of people trust AI companies, trust becomes something you can compete on. A business that can show how it uses AI, what it does with data, and where a human stays in the loop will win work that a less careful competitor loses.
Staff who fear being replaced will quietly resist a tool they believe is there to cut headcount.
Customers who distrust AI vendors will read a clumsy chatbot as a sign you have stopped caring.
Regulators and large buyers increasingly ask how a supplier governs AI before they sign anything.
Each of these is a cost if you ignore it. Handled well, each is also an opening to look more trustworthy than the competition, which is rare ground in a market full of overblown AI claims.
How to roll out AI without spooking people
The survey points to a rollout that leads with transparency and keeps people in control. The steps are not complicated, but they do need to be deliberate.
Tell staff plainly what the AI will and will not do, and frame it as removing dull work rather than removing people.
Keep a person reviewing anything that reaches a customer or carries legal weight.
Be open with customers when they are dealing with an AI system, and make it easy to reach a human.
Handle personal data to the standard the Privacy Act expects, and write down how you do it.
Choose vendors on their safety record, not only their benchmark scores.
A modest first project, scoped to one painful task and run with these guardrails, typically costs an Australian SMB around $40,000. It earns trust as much as it saves time, and the trust is the part that compounds over the projects that follow.
Where Claude fits
We build on Claude, from Anthropic, partly because safety and transparency are central to how Anthropic positions the model. That posture is easier to stand behind when a customer or a regulator asks how your AI behaves. It is not a magic answer. The governance, the human review, and the honest communication are still your job. But starting from a model with a careful safety story makes the trust conversation shorter. For a small Australian firm without a dedicated AI team, that shorter conversation is worth real money, because it shortens sales cycles and reduces the back and forth with cautious buyers. If you want help planning an AI rollout that your staff and customers actually accept, book a brainstorm.
