The Real AI Ethics Question for Small Business

The big AI question gets asked in a dramatic way: will AI kill humanity?

That sounds like science fiction, but the conversation is no longer only happening in internet forums or conspiracy corners. Serious researchers, frontier AI leaders, policy groups, and governments are openly discussing whether advanced AI could create risks on a global scale.

That question matters, but it is not the only question a small business owner should be asking.

You are probably not building the next frontier AI model. You are not deciding how global AI systems are trained, released, governed, or restricted. But you may already be using AI inside your business for emails, customer replies, sales documents, meeting summaries, research, reporting, hiring, marketing, operations, or decision support.

That makes the ethics question practical.

The real question is not only whether AI becomes dangerous someday. It is whether you are using it carelessly right now.

AI Risk Sounds Big, but It Starts Small

When people talk about AI risk, the conversation often jumps straight to extinction, superintelligence, robots, or global collapse. That is one layer of the issue, and it should not be ignored.

But for most small businesses, the immediate risk is closer to home.

It is the customer data pasted into a public AI tool. It is the staff member using an unsanctioned AI app to handle sensitive information. It is the AI-drafted email that goes out without proper review. It is the sales proposal with the wrong promise. It is the customer support answer that sounds confident but is wrong.

These are not science fiction problems.

They are business problems.

The moment AI starts touching real business information, ethics stops being abstract. It becomes about access, data, responsibility, review, accuracy, and trust.

You Cannot Control the Frontier Models

There are legitimate concerns about how powerful AI models are being developed. People like Geoffrey Hinton, Dario Amodei, Elon Musk, and other AI researchers have publicly expressed concern that advanced AI could create serious risks if it is not governed well.

Those warnings matter, but they are not something a small business owner can directly control.

You cannot personally decide how OpenAI, Google, Anthropic, xAI, Meta, or any other frontier lab trains and releases its models. You cannot set global AI policy by yourself. You cannot stop every bad actor from using AI badly.

What you can control is how AI is used inside your business.

That is where your responsibility starts.

The Practical AI Ethics Question

The practical AI ethics question for a small business is simple:

What are you letting AI do, what information are you giving it, and who is responsible for the result?

If AI drafts a customer reply, who checks it?

If AI summarises a meeting, who confirms the actions are correct?

If AI writes a proposal, who checks the scope and pricing?

If AI helps with hiring, who checks for bias?

If AI reviews customer data, who decided that data was safe to use?

If AI connects to your inbox, Drive, CRM, calendar, or finance tools, who understands what it can access?

That is the real work.

Not panic. Not avoidance. Not blind enthusiasm. Responsible use.

Do Not Let Fear Stop You From Using AI

There is a lot of opportunity in AI. Used properly, it can help small businesses move faster, make better decisions, reduce admin, improve documentation, support customer service, and give owners more capacity.

Avoiding AI completely because the big-picture risks feel uncomfortable may leave your business behind.

The better approach is to use AI with rules.

This is not very different from other technology. Email created new risks. Cloud storage created new risks. Online payments created new risks. Social media created new risks. Businesses still use them, but they need policies, controls, access rules, and good judgment.

AI is the same, but faster and more powerful.

That means you should not ignore it, and you should not casually throw it into every part of the business without thinking.

The Risk Is Not Only That AI Gets Too Powerful

For small businesses, the bigger everyday risk is that humans get too casual.

AI is convenient. That is the point. It can write faster than you, summarise faster than you, compare more information than you, and produce a clean-looking output in seconds.

That speed can make people lazy.

They stop checking. They stop asking where the information came from. They stop reading the output properly. They copy and paste customer information without thinking. They assume the AI knows what it is doing because the answer sounds polished.

That is where problems start.

The danger is not only that AI becomes too powerful one day. It is that a business owner or employee trusts it too quickly today.

Customer Data Needs Clear Rules

One of the first areas to control is customer data.

Small businesses often hold more sensitive information than they realise. Emails, call notes, payment details, addresses, contracts, support tickets, health information, financial information, staff information, and internal decision-making can all end up inside business systems.

Before putting any of that into AI, ask whether it should be there.

Not all AI tools have the same privacy terms. Not all tools are approved for business use. Not all staff understand the difference between using AI for a rough writing task and uploading private customer records into a third-party system.

You need rules that explain what can and cannot be entered into AI tools.

For example, a business may allow AI to rewrite a generic marketing paragraph, but not allow staff to paste full customer records, medical details, bank information, passwords, contracts, or confidential internal documents into unapproved tools.

That line should be written down.

Your Inbox Is Not Just an Inbox

Connecting AI to Gmail, Outlook, or any business inbox can be useful. AI can summarise email threads, sort messages, draft replies, identify priorities, and reduce time spent in the inbox.

But your inbox is one of the most sensitive systems in your business.

It may contain customer details, login links, verification codes, invoices, legal conversations, HR information, pricing discussions, bank alerts, and private client context. Giving AI access to that inbox is not the same as asking it to rewrite a paragraph.

That does not mean you should never connect AI to email. It means you should understand exactly what access you are granting.

What can the AI read? Can it send emails? Can it delete messages? Can it create drafts? Can it access attachments? Who reviews the output? What happens if it sends or suggests the wrong thing?

Those questions need answers before the connection is made.

SixFive’s Google Workspace services can help small businesses structure accounts, access, shared drives, admin roles, and security settings before layering more AI tools on top.

Turn Off Training Where Appropriate

Most AI tools have settings that affect whether your prompts, uploads, and outputs may be used to improve or train their systems.

Business owners should check those settings.

This is a simple but important step. If you are using AI for business work, go into the settings for the tools you use and review data controls, training settings, privacy options, and workspace-level admin controls.

Do not assume the defaults match your risk tolerance.

Even if the setting is not a complete solution, it is part of responsible use. It shows that the business has thought about how its data is handled.

For a broader check of access, passwords, devices, data, email security, and backup, start with the Small Business Cyber Profile.

Your Team May Already Be Using AI

Even if the business has not officially adopted AI, your staff may already be using it.

That is normal. People want to get their work done. If they are stuck, under pressure, or unsure how to complete a task, they may open ChatGPT, Gemini, Claude, Perplexity, or another AI tool and use it without asking.

That is shadow AI.

It is similar to shadow IT, where employees sign up for tools the business does not know about. The difference is that with AI, the employee may be pasting in business data, customer information, internal documents, or strategic material.

This is why banning AI rarely works. People will still use it if it helps them.

The better answer is to give them approved tools, basic rules, and clear guidance.

Tell staff what AI tools they can use, what they can use them for, what information must not be entered, what outputs need review, and when they should ask before connecting a tool to business systems.

AI Needs Role-Based Responsibility

As AI becomes part of business workflows, it should be included in role responsibilities.

If a staff member uses AI for customer support drafts, they are responsible for checking those drafts before anything is sent. If someone uses AI to prepare reports, they are responsible for checking the numbers, context, and conclusions. If a manager uses AI in hiring, they are responsible for making sure it does not create unfair decisions.

This should not be vague.

Just as each person has responsibility for specific tools, systems, or processes, they should also understand which AI-assisted workflows they own.

That might include checking outputs, keeping source documents up to date, reporting errors, improving prompts, reviewing automations, and making sure customer-facing content is approved.

AI does not remove responsibility from the role. It changes how the role is performed.

Keep a Human in the Loop

Human review is one of the most important parts of responsible AI use.

If AI is producing anything that could affect a customer, employee, supplier, prospect, policy, contract, price, public claim, or business decision, a human should review it before it becomes final.

This is especially important for customer replies, proposals, pricing explanations, hiring documents, internal policies, legal language, compliance language, financial summaries, marketing claims, technical support answers, and health-related or finance-related content.

AI can help prepare the work.

It should not automatically own the final decision.

A human in the loop does not mean the business is avoiding AI. It means the business is using AI with judgment.

AI Can Make Things Up

AI can sound confident and still be wrong.

It can invent citations, misunderstand data, make assumptions, misread context, overstate claims, flatten nuance, or present a biased answer as if it is neutral. It has improved, but it still makes mistakes.

This matters most when the business uses AI for research, customer advice, public claims, policy documents, financial summaries, or technical explanations.

Do not treat an AI answer as truth just because it sounds clean.

Ask where the information came from. Check the source. Compare against trusted references. Ask for alternative views. Use tools that show citations or source grounding where possible.

Google’s Gemini Notebook is useful for this kind of work because it is designed to work from sources you provide, rather than relying only on a general AI response. It can help you analyse documents, notes, websites, and other material while keeping the answer grounded in sources you choose.

Even then, you still need to check the output.

Use AI Inside a Defined Box

One useful way to reduce AI risk is to narrow the box.

Instead of asking a general AI tool to answer from the whole internet, give it approved material to work from. That could be your own policy documents, product information, service pages, FAQs, previous proposals, meeting transcripts, or research papers you have selected.

This does two things.

First, it improves quality because the AI has better context.

Second, it reduces risk because the AI is less likely to pull from random or unsuitable material.

This does not make it perfect. AI can still misunderstand the documents. But it gives the business more control than asking a general model to work everything out from scratch.

The more important the output, the more important source control becomes.

Be Careful With AI in Hiring

Hiring is one area where AI needs extra care.

AI can help organise job descriptions, summarise candidate notes, create interview questions, and structure evaluation forms. That can be useful.

But using AI to screen applicants or rank people can introduce serious ethical and legal risks if the model reflects bias, uses poor data, or makes decisions in ways the business cannot explain.

If AI is used in hiring, the business should know exactly how it is being used, what data it is looking at, who checks the result, and whether the process treats candidates fairly.

Do not let AI quietly become the reason someone is rejected without human oversight.

That is not just an AI issue. It is a people issue.

Be Careful With AI in Finance and Health

Financial and health-related information also needs a higher bar.

If your business handles medical, wellbeing, insurance, aged care, disability, tax, accounting, legal, lending, investment, or financial advice, AI use should be tightly controlled.

AI can help draft notes, summarise information, structure a report, or prepare questions. But any content that influences someone’s health, finances, legal position, or major life decision needs proper human expertise and review.

This is where businesses can get into real trouble if they confuse speed with safety.

A polished AI answer can still be incomplete, inappropriate, or wrong for the individual situation.

Do Not Let AI Make Unsupported Claims

Marketing is another area where AI can create problems.

AI is very good at writing persuasive copy. It can also overstate benefits, invent proof, make claims the business cannot support, or turn a mild advantage into an exaggerated promise.

That is a risk.

If AI writes website copy, ads, email campaigns, landing pages, proposals, or product descriptions, someone needs to check the claims. Can you prove what it says? Is the promise accurate? Is the comparison fair? Does it match what your service actually delivers?

The Federal Trade Commission’s business guidance says advertising claims should be truthful, not deceptive or unfair, and evidence-based. That principle applies whether the words came from a human or AI.

The business is still responsible for what it publishes.

Create an AI Use Policy

Every small business using AI should have a simple AI use policy.

It does not need to be twenty pages. It needs to be clear enough that staff know what is allowed.

A basic policy should cover approved AI tools, prohibited data, sensitive information, customer data, use in hiring, use in finance or health-related work, human review requirements, output checking, source requirements, disclosure rules, and what to do when AI gets something wrong.

This gives staff a way to use AI without guessing.

It also gives the business a way to manage risk without stopping everyone from using useful tools.

If your business does not yet have clear workflows or documentation, SixFive’s Notion SOP template can help you start writing down repeatable processes before AI gets layered on top.

Ask Better Questions Before Using AI

Before using AI in any workflow, ask a few practical questions.

What is AI being used for?

What data does it need?

Is the data sensitive?

Which tool is being used?

Is that tool approved?

Can the AI access email, files, calendars, CRM, finance systems, or customer records?

Who checks the output?

Could this affect a customer, employee, supplier, or business decision?

What happens if the AI is wrong?

Should the customer or staff member know AI was used?

If you cannot answer those questions, the workflow is not ready yet.

Use a Simple Risk Framework

You do not need to become an AI governance expert to start using AI responsibly. But it helps to think in a simple framework.

The NIST AI Risk Management Framework uses four broad ideas: govern, map, measure, and manage.

In small business language, that means:

Govern: decide the rules.

Map: understand where AI is being used and what it affects.

Measure: check whether it is working and where the risks are.

Manage: reduce the risks, improve the workflow, and keep reviewing it.

That is a practical way to think about AI without turning it into a huge compliance project.

Start small. Pick the AI workflows you already use. Then apply the same thinking.

AI Ethics Is Not Just for Big Companies

Small businesses sometimes assume AI governance is only for large enterprises, government, or regulated industries.

That is not true.

If AI touches your customers, staff, data, decisions, or public communication, ethics matters.

A small business may not need a full governance department, but it still needs basic rules. It still needs human review. It still needs to protect customer information. It still needs to avoid unsupported claims. It still needs to understand what tools staff are using.

Responsible AI does not have to be complicated. It just has to be intentional.

The Bottom Line

The big AI question may be whether AI could become dangerous at a global scale.

The small business question is more immediate: are you using AI responsibly right now?

AI can be a force multiplier. It can help you write, summarise, research, report, plan, and make decisions faster. But it also creates risk if you use it casually, paste in sensitive data, trust outputs without checking, connect it to systems without understanding the access, or let staff use random tools without rules.

You do not need to panic. You do need to take responsibility.

The risk is not only that AI becomes too powerful. It is that humans become too casual with a powerful tool.

What to Do Next

Pick one AI workflow your business already uses. Review what data it touches, what tool is being used, who checks the output, and what could happen if the answer is wrong.

Then write down the rule. Keep it simple. What can AI do? What can it not do? What information must never go into it? Who must review the output? Where does the approved version live?

If you want to check where your business is exposed, start with the AI Readiness Audit. It will help you review whether your systems, documentation, data, and rules are ready for responsible AI use.

You can also take the Small Business Cyber Profile if your concern is access, passwords, devices, data, email security, and backup.

For practical support mapping your tools, workflows, permissions, and AI risk, review SixFive’s Digital Roadmap or book an appointment with SixFive.