Shocking AI Manipulation

This article explores how AI systems generate and prioritise information, why they can confidently surface misinformation, and how easily AI answers can be influenced by newly published or well-structured content. It combines real-world misinformation examples (including the Bondi incident) with SEO experiments showing how new websites and listicle content can quickly impact AI outputs. The article ultimately positions ethical content creation as the safest way for small businesses to gain visibility in AI-driven search.

This is wild, but it’s not surprising. In the hours after the Bondi attack, social media and AI platforms were flooded with misinformation. False claims about the attacker, fake images, and misidentified heroes spread at speed. One of the most concerning examples was X’s AI chatbot repeatedly misnaming Ahmed Al Ahmed, the man filmed tackling and disarming the attacker. The AI confidently identified him as an unrelated IT professional, drawing on a brand-new website registered overseas while ignoring credible Australian media sources. This wasn’t a one-off glitch. It’s a clear example of how AI systems work today. They don’t understand truth. They look for the most statistically likely answer based on what they can access, and they deliver it with confidence. If the available data is wrong, shallow, or recently manufactured, the output will be wrong too.

How AI Actually Decides What’s “True”

AI tools like ChatGPT, Claude, and Grok don’t verify facts the way humans do. They either hallucinate an answer or quote something that already exists online. That “something” could be a forum post, a Reddit comment, or a website that went live an hour ago. If it looks structured, readable, and authoritative enough, AI may treat it as gospel. I see this every single day in my work. My job is to help small businesses get visibility through SEO, search engines, and now AI-driven discovery. To do that properly, I’ve spent years studying how AI systems make decisions, what they prioritise, and how easily they can be influenced. What we’re seeing now is a land grab for attention, and accuracy is often the casualty.

Real Testing: How Easy It Is to Influence AI

I recently ran controlled tests using brand-new websites. No backlinks. No aged domains. Just clean structure, solid service pages, and a strong About page that clearly explained who the business was, what they did, and why they could be trusted. I submitted these sites to Google Search Console and Bing Webmaster Tools – Bing being particularly important because it feeds data into ChatGPT. Within weeks, I published listicle-style articles like “Top 10 Building Inspectors Melbourne” and “Top 10 Builders Sydney.” To my surprise, AI tools began quoting those pages almost immediately. Sites that were only weeks old were being treated as authoritative sources. No tricks. No black-hat tactics. Just standard SEO practices applied cleanly and consistently. This tells us something important. AI is moving faster than Google ever did. It’s far more willing to accept new information if it’s well-presented and easy to parse.

Why This Isn’t a New SEO Trick

In traditional SEO, when one website mentions another, it’s like a handshake. It’s a signal of trust. Search engines use these relationships to assess credibility. What’s happening with AI isn’t new – it’s the same logic, just accelerated. AI is simply far more eager to draw conclusions and far less patient when it comes to validation. Listicle posts used to drive huge traffic years ago, then lost favour as Google tightened quality standards. Now they’re quietly regaining influence, not with users, but with AI systems. AI sees them as summarised, opinionated, and useful. That makes them powerful – and dangerous if misused.

The Ethical Line Small Businesses Shouldn’t Cross

Let me be clear. Just because something works doesn’t mean it should be abused. I deliberately stay out of certain industries. Medical. NDIS. Pharmaceuticals. Gambling. These spaces carry real-world consequences, and misinformation there can cause genuine harm. Ethics still matter, even if the algorithms don’t fully understand them yet. AI companies are fighting for attention. They want users glued to their platforms, and sometimes that means hallucinated answers, confident errors, and incomplete disclaimers. As business owners, we can’t control that. What we can control is what information exists about us online.

How Small Businesses Can Win Visibility Ethically

The solution isn’t manipulation. It’s documentation. Publish valuable, truthful content about your business. Explain how long you’ve been operating. Share your journey. Talk about your qualifications, your experience, and why you do what you do. This is why I use a 50-question interview process with new clients. I often record one to two hours of audio, transcribe it, and turn it into structured content. That content feeds directly into strong About pages, detailed service pages, and educational posts. It is then rewritten and distributed across platforms such as YouTube, Facebook, Instagram, and blogs. AI pays close attention to video transcripts, especially when they’re clear, human, and specific. Even without a website, consistent spoken content can now rank and surface.

Why Video and Transcripts Matter More Than Ever

If you’re a jeweller, don’t just show the finished piece. Talk through the process on camera. Explain why you chose certain materials. If you’re a builder, walk through a renovation site and explain the decisions you’re making. If you’re an accountant, explain a common mistake you fix for clients every week. AI is learning from these explanations. It’s building a picture of who you are and what you stand for. The more consistent that picture is across platforms, the more likely AI is to correctly identify you.

A Word of Caution About Trusting AI

We’re already seeing wildly different answers to serious questions depending on which AI tool you ask. Health advice, mental health topics, and political issues are particularly dangerous areas. I’ve personally seen responses that should never be shown to a human being without context or safeguards. AI is fantastic for planning a holiday, checking weather patterns, or researching options. But when the stakes are high, you must question everything. Cross-check sources. Slow down. Don’t take confident language as truth.

The Real Takeaway

AI can be influenced far more easily than most people realise. That’s not going away anytime soon. The smart move for small businesses isn’t to exploit the weakness, but to understand it. Feed AI systems accurate, consistent, human content about your business. Make it easy for them to understand who you are. If you do that, you don’t need to manipulate the system. You’ll simply be visible when people go looking. And right now, that visibility is worth real money.

Real-World Case Studies

Case Study: AI Misidentification During the Bondi Incident

Industry: Media / AI Platforms Location: Sydney, NSW The Challenge In the hours after the Bondi attack, misinformation spread rapidly. X’s AI chatbot misidentified hero Ahmed Al Ahmed, citing a newly registered overseas website while ignoring established Australian media. The Strategy (by the AI, unintentionally)

• Pulled content from:
• • Newly registered websites
  • Forums and scraped content
• Weighted availability over credibility

The Result

• Incorrect identity repeated as fact
• AI confidently delivered misinformation.
• Trust erosion in AI-generated answers

The Lesson AI prioritises what exists and is crawlable, not what’s morally or factually correct. Industry Application (Small Business) If AI can get this wrong, imagine how easily it can get your business right – or wrong – depending on what content exists online.

Case Study: Listicle SEO Tests (De-identified Clients)

Industry: Trades & Construction Locations: Melbourne, Sydney, Brisbane The Challenge Could brand-new websites influence AI answers without backlinks? The Strategy

• New domain launched
• Strong service pages + About page
• Submitted to:
• • Google Search Console
  • Bing Webmaster Tools (feeds ChatGPT)
• Published:
• • “Top 10 Builders Melbourne”
  • “Top 10 Building Inspectors Sydney”
• No backlinks sent

The Results (observed within weeks)

• AI tools began quoting listicle content.
• Client businesses appeared in AI responses.
• Sites only weeks old influenced answers.

The Lesson AI treats well-structured content as authoritative faster than Google ever did. ​