'AI Is Going to Replace Us All'
The common topic I encounter in all audiences. A scary thought, but something that I think is further away than most may realize.
The mainstream rise of AI is another one of those "technology landmark" moments. Before AI there was the internet, and before the internet there were phones. Each of these shifts were dramatic and even frightening at the time, but they built the world as we know it today.
When the telephone arrived, operators feared obsolescence. When the internet emerged, entire industries braced for collapse. Jobs were lost; but far more were created. The internet didn't just automate existing work, it invented entirely new categories of work that nobody had imagined before. AI will do the same.
AI: The Modern Intern
- Let's be fair. AI already has some strengths:
- Executing basic and repetitive code
- Gathering and summarizing searchable information
- Generating quick presentations and reports
- Pattern recognition and data analysis
- Drafting first-pass content at speed
These capabilities are real and accelerating, but everything produced still always needs to be edited and reviewed. If your job consists entirely of one of these tasks, you should absolutely be paying attention. Otherwise, for most professionals, AI is better understood as a very fast, very confident intern. I personally call today's AI tools "glorified search engines" depending on what you use them for. The issue with overconfident interns is they often make things up.
Why There Will Always Have to Be a Supervisor
The truth: AI is often wrong, especially with multilayered tasks. The consequences of being wrong at scale are too large for most organizations to absorb without a human in the loop.
There are three reasons human oversight isn't going away anytime soon.
It gets things wrong. AI systems hallucinate: they generate plausible-sounding information that is simply not true. This isn't an edge case. It happens regularly across industries, and the cost of catching those errors falls entirely on humans.
Someone has to be accountable. Organizations need a human who can be held legally and professionally responsible for decisions. AI cannot be sued, licensed, fired, or deposed. When something goes wrong, not if, there has to be a person at the end of the chain who owns it. It is reasonable to assign an individual responsible for the information chosen to be used from AI.
It has bias, which lacks judgment. AI models are built from preexisting data and rule sets developed by engineers, meaning they inherit those biases and cannot truly understand organizational context, ethical nuance, or the kind of ambiguous, high-pressure decision-making that defines upper-level roles. A CEO deciding whether to pull the plug on a program is not just entering a prompt. They are making a judgment call built on years of experience, relationships, and institutional knowledge that no model can genuinely mimic.
When AI Gets It Wrong: Real Examples
AI causing serious problems without adequate supervision is not new. These aren't fringe cases: they span consulting, legal, and technology.
Consulting
Deloitte — The Phantom Government Report
In July 2025, Deloitte published a 237-page report for Australia's Department of Employment and Workplace Relations: a contract worth approximately $290,000 USD.[1] Within weeks of publication, Dr. Chris Rudge, a researcher at Sydney Law School, flagged that the report was riddled with fabricated citations, references to nonexistent academic papers, and an invented quote attributed to a federal court judge, cataloguing roughly 20 errors in total.[2]
A corrected version was quietly uploaded in October 2025, disclosing that Microsoft's Azure OpenAI GPT-4o had been used to help draft sections of the report. Deloitte agreed to refund the final installment of the contract fee.[3] The core problem: AI filled gaps with confident fiction, and nobody caught it before publication.
Legal
MyPillow's Lawyers — Fined in Federal Court
In February 2025, attorneys Christopher Kachouroff and Jennifer DeMaster, representing MyPillow CEO Mike Lindell in a Colorado defamation case, filed a motion containing nearly 30 defective citations including references to cases that simply do not exist.[4]
On July 7, 2025, U.S. District Judge Nina Y. Wang fined each attorney $3,000, noting the brief had been drafted with AI assistance and was not adequately reviewed before filing. The errors ranged from missing basic court information to citing entirely fabricated case law.[5] This echoed a 2023 case that legal circles dubbed a "teachable moment." Apparently not everyone learned the lesson.
Technology
Wolf River Electric vs. Google
Wolf River Electric, an Isanti, Minnesota solar company, filed a defamation lawsuit against Google in March 2025 after its AI Overview feature falsely stated the company was facing a lawsuit from the Minnesota Attorney General for deceptive business practices.[6] The company was not a defendant in any such case, the AG had sued four solar lending companies, none of them Wolf River.[7]
By the time the error was caught, multiple customers had canceled contracts. The complaint documents specific losses: a $150,000 contract terminated on March 5, 2025; a nonprofit pulling $174,044 in projects citing the fabricated lawsuits; an estimated $24.7 million in losses from 2024 alone.[8] Wolf River is seeking between $110 million and $210 million in damages. Google has denied responsibility.
These cases span three entirely different domains, which tells you something important: this is not a niche problem. It is a structural one.
The Cybersecurity Angle
For those of us in security, the picture is even more nuanced.
AI-assisted threat detection is real and genuinely useful. Models can process at a speed and scale no human analyst can match. But the assumption that AI can replace proven security tooling and human judgment has already started to cause problems.
What we continue to see is warnings from other professionals and basic logic being forgotten.
Meanwhile, on the offensive side, AI is already being used to accelerate attacks: enabling less-skilled actors to launch more sophisticated phishing campaigns, generate convincing social engineering content, and automate reconnaissance. The threat is evolving faster than most defenses.
This creates an arms race. As the field expands for attackers and defenders alike, AI in cybersecurity demands more human expertise, not less. Someone has to evaluate the outputs, make the calls, understand what the models are getting wrong.
The Real Question
The fear that "AI will replace us all" is fair, but it's the wrong mindset. The right question is: will people who are proficient in AI replace people who are not?
With how AI is advancing, it makes the most sense to shift perspective from passive dread to active preparation. AI is a tool. It is powerful, fast, and sometimes dangerously overconfident, but just a tool in our box of skill sets. Like every technology landmark before it, the people who learn to work alongside it, supervise it, and catch its mistakes will be the ones who come out ahead.
The jobs that disappear will be the ones where humans were already acting like machines: executing repetitive tasks without judgment, accountability, or context. The jobs that grow will be the ones that require exactly what AI cannot provide.
Stay curious, stay secure.
Sources
- Fortune, Nino Paoli, "Deloitte was caught using AI in $290,000 report to help the Australian government crack down on welfare after a researcher flagged hallucinations," October 7, 2025. fortune.com
- Above the Law, "Law Professor Catches Deloitte Using Made-Up AI Hallucinations In Government Report," October 8, 2025. abovethelaw.com
- CFO Dive, "Deloitte AI debacle seen as wake-up call for corporate finance," October 14, 2025. cfodive.com
- NPR, "A recent high-profile case of AI hallucination serves as a stark warning," July 10, 2025. npr.org
- Colorado Sun, "MyPillow CEO's lawyers fined for AI-generated court filing in Denver defamation case," July 8, 2025. coloradosun.com
- GovTech / Minnesota Star Tribune, "Minnesota Solar Company Sues Google Over AI Summary," June 13, 2025. govtech.com
- Reason / Volokh Conspiracy, "Large Libel Models: Small Business Sues Google, Claiming AI Overview in Searches Hallucinated Attorney General Lawsuit," June 11, 2025. reason.com
- Medium / Purdy House Publishing, "An AI Defamed a Solar Company. Google Says It's Not Responsible," April 2026. medium.com
- Jake Williams (MalwareJake) professional profile — IANS Research Faculty, VP R&D at Hunter Strategy. iansresearch.com
Member discussion