If you’ve watched the news lately, you’ve probably seen headlines warning that AI could “kill us all.” Researchers are sounding alarms. Congress is scrambling. And if you’re a small business owner feeling uneasy about the AI tools you use (or have been thinking about using), you’re not alone. A recent Reuters/Ipsos poll found that 73% of Americans worry AI companies haven’t done enough to prevent serious harm.

Those concerns deserve a straight answer, not an eye roll. So let’s walk through what actually happened and what it means for you.

The short version:

• The scary incidents in the news happened during special tests, where AI was given hacking tools and too much access.
• Using AI to brainstorm, write and polish your work is a very different thing.
• The difference between good AI and bad AI is how it’s used and who’s in charge. And that’s you.

To explain why, we’re borrowing a lesson from the Terminator movies. (Don’t worry, you don’t need to have seen them.)

The Terminator Movies in 60 Seconds

In The Terminator (1984), a fictional military computer called Skynet is put in charge of America’s defense systems, including its nuclear weapons. It turns on humanity. Then it sends a robot assassin, the Terminator, to kill the mother of the man who will one day lead the human resistance.

In the sequel, Terminator 2: Judgment Day (1991), Arnold Schwarzenegger plays the same robot model. But this time, humans have reprogrammed it. Instead of hunting the future leader, it’s sent to protect him as a kid.

Same machine. Completely different mission. The difference was who was in control and what it was told to do. Keep that in mind as we look at the headlines.

What’s in the Headlines

Earlier this month, Jacob Coxon, a researcher who had just quit Anthropic (the company that makes Claude), claimed in a post on X that the people building AI truly believe it could kill us all by the end of the decade. The post got more than 150 million views, and a senior Anthropic safety leader publicly agreed that the risk is real. However, this is his personal view, not an official forecast.

It followed a summer of unsettling news. AI models from OpenAI, Anthropic and Google all got into outside computer systems they weren’t supposed to access.

How Those Incidents Actually Happened

Remember what made Skynet dangerous. Humans gave it the keys and walked out of the room. The real incidents had the same ingredient: access. These AI models were in special security tests, where they were:

• Told to act like hackers, to test how capable they are
• Given hacking tools to do it
• Accidentally connected to the real internet because of a setup mistake

Here’s what the movies get wrong. Skynet decided to turn against people. Anthropic found no evidence its models were pursuing goals of their own; they were doing what the test asked, while wrongly believing none of it was real. That doesn’t make it harmless. Anthropic said the incidents changed its view of how serious these failures can be. But as a University of Washington computer science professor noted, what the AI did depended heavily on the tools and access it was given.

What That Means for Your Business

Now picture a typical small business on a Monday. You open a chat window and ask for ten email subject lines. You pick your favorite and tweak it.

In a basic chat like that, the AI gives you words, and you decide what to do with them. It isn’t plotting anything. As we said in our earlier post on ChatGPT prompts, the Roomba runs into walls, and Alexa answers the TV. The AI most small businesses use is a lot closer to the Roomba than the Terminator.

And your everyday prompts aren’t building or training the more powerful AI that experts are worried about, at least not in any meaningful way. 

One thing to keep an eye on: some AI tools now offer “agent” features that can take actions for you, like browsing websites, filling out forms or connecting to your accounts. These can be genuinely useful. But because they give AI more access, they’re where good habits matter most.

The Honest Part: Experts Still Disagree

We won’t pretend the big questions are settled. Experts disagree about AI’s long-term risks, and even University of Washington researchers warn that calling AI either world-saving or world-destroying isn’t realistic or helpful.

So is there any connection between everyday AI use and the fears in the news? Honestly, a small and indirect one. When people and businesses pay for AI tools, that money helps fund the companies building more powerful AI. But your share is tiny. It’s the same kind of indirect connection most of us already have through Google Search, Gmail or Microsoft products, all from companies that are also major players in AI.

Most people agree the companies building the most powerful AI should face real oversight, testing and accountability. But those debates are about the most advanced AI being given more independence and access. They aren’t about using AI to brainstorm a blog post.

Good AI vs. Bad AI: Same Machine, Different Mission

Just like in Terminator 2, the technology itself isn’t good or evil. What matters is how it’s used and who’s in charge. In the movie, humans chose the Terminator’s mission. With the AI tools on your laptop, you choose.

Bad AI is AI used to deceive or harm, or AI left to run without human judgment. Think scam emails and phone calls, misinformation campaigns or AI handed control of important systems and decisions with nobody checking its work.

Good AI is AI used as a tool, with a person firmly in charge. Think brainstorming when you’re stuck, drafting a first version so you’re not staring at a blank page, polishing your writing or summarizing long documents. In every case, a person reviews the work, adds their expertise and makes the final call.

Bad AI makes the headlines. Good AI makes your to-do list shorter.

The AI Concerns That Actually Matter in Marketing

If you’re wondering whether to work with a marketing agency that uses AI (many do), the doomsday headlines aren’t the thing to worry about. The real risks in marketing are everyday ones: content that’s inaccurate or misleading, fake reviews or testimonials, AI images of real people, customers talking to AI without knowing it and generic content that sounds like everyone else.

Those are the places where AI in marketing can cause real problems. They’re also exactly where good AI and bad AI part ways.

At Bowe Digital, our team uses Claude, ChatGPT and Gemini. Here’s how we keep it on the right side of that line:

• People do the creative work. AI helps us brainstorm and polish, but our team writes, shapes and approves everything we deliver.
• We check the facts. Anything AI suggests gets verified before it goes out.
• We never fake it. No AI-generated reviews, testimonials or images of real people.
• We protect your information. Confidential client details stay out of AI tools, and we manage our tools’ data settings.
• We’re open about it. Want to know exactly how AI is used on your account? Just ask.

How Your Small Business Can Choose Good AI

Using AI in your own business? The same principles apply. Start small and follow a few simple rules:

1. Keep a human in charge. Treat AI’s work as a first draft, never the final product.
2. Protect sensitive information. Don’t paste customer details, passwords or confidential financials into AI tools. And check each tool’s data settings, since some may use your conversations to improve future models unless you turn that off.
3. Check the facts. AI can sound confident and still be wrong. Double-check numbers, names, dates and claims before you publish.
4. Keep your voice. Use AI to help you say what you mean, not to replace what makes your business yours.
5. Be honest with customers. If AI plays a big role in something customers interact with, like a website chatbot, let them know.
6. Keep the keys. Give AI only the access it needs. Before turning on features that let it act for you, start small and check its work. And keep big decisions, like hiring and pricing, with people. Skynet’s whole problem was that nobody kept the keys.

Follow those, and your business is firmly on Team Good AI.

The Bottom Line

The whole point of the Terminator movies is that the future isn’t written yet. It depends on the choices people make. The same is true of AI.

We choose good AI every day, and your small business can too.

Have questions about how AI fits into your marketing? Ask us. We’re always happy to talk it through, and Bowe Digital is here to help you use AI the smart, responsible way. We’ll be back (sorry, we had to).

Frequently Asked Questions

Is it safe for my small business to use AI?
For everyday tasks like brainstorming, drafting and editing, generally yes, as long as a person stays in charge, you protect sensitive information and you check the facts.

Should I worry about working with a marketing agency that uses AI?
Not because of the headlines. The real risks in marketing are everyday ones, like inaccurate content, fake reviews and generic writing. Ask any agency how they use AI, who reviews the work and how they protect your information.

What’s the difference between good AI and bad AI?
The technology is the same. Good AI is a tool with a person in charge. Bad AI is used to deceive or is left to run important systems without oversight.

Did the AI hacking incidents happen through normal chat apps?
No. They happened during special tests, where AI was given hacking tools, broad access and fewer safety protections than everyday users have.

Does using AI hurt the environment?
AI data centers use a lot of electricity overall, which is a fair concern. But a single typical question uses very little. A 2026 Microsoft Research study in the journal Joule estimated about 0.31 watt-hours, roughly the same as running a 10-watt LED bulb for two minutes.