AI Can Get You Fired and Bankrupt
Key Highlights
- AI hallucinations can produce false information that undermines workplace safety assessments and emergency plans, leading to legal liabilities.
- Using AI in safety planning without proper review can result in violations of OSHA standards and significant jury awards in lawsuits.
- Federal courts and agencies are increasingly scrutinizing AI-generated work, with documented cases of hallucinations and errors leading to sanctions and fines.
- Employers should follow recognized safety standards, prioritize thorough training, and avoid over-reliance on AI to ensure compliance and safety.
- Legal cases have demonstrated that AI errors can be used against employers in court, potentially causing bankruptcy and reputational damage.
Pundits predict the AI bubble could burst and be worthy of 1929 or 2008—or there can be an orderly adoption of AI at every workplace. Either way, there are serious issues with how you use AI in our workplace security and safety world.
I'm coming to you today to repeat and amplify this warning because many employers aren’t listening.
Workplace security, safety and emergency management are complex issues requiring insight and knowledge. Fans of AI see it as a fast, easy way to remove complexity and save time. The hallucination: AI will create your workplace security assessment today; your Emergency Action Plan tomorrow; and your people trained the next day. Done!
I have served as an expert consultant and witness in court from California to Florida; from Texas to Connecticut. I can directly report that, when you’re sued after any individual was injured or killed at your workplace, all your assessments, plans, training materials, drills and exercises will be introduced as evidence at court. This is to establish whether you have met your duty of care to keep all personnel safe at your workplace: “Each employer shall furnish …a place of employment …free from recognized hazards that are causing or are likely to cause death or serious physical harm to employees.” [Federal law passed in 1970: 29 U.S.C. § 654(a)(1)].
Litigators will endeavor to establish your failure to plan and failure to train. They’ll challenge the quality and scope of your assessment, planning, your training materials, and your sign-in sheets. Most of you do not have an OSHA-compliant Emergency Action Plan, thereby failing to meet your duty of care. Once your failure is established in front of a jury, you will be facing huge jury awards. If you have used AI in your planning, litigators will move to multiply that dollar award.
I can cite jury awards of $5 million to $1 billion in the last two years against employers like you who experienced a workplace injury or death. Most employers could not survive awards of that magnitude.
Example #1: This AI hallucination had been tried in front of a federal judge in New York City. Motions were made that looked strange to the judge. An investigation found that these lawyers had used AI. The judge determined that the facts were invented. The sources for the facts were invented. The precedents cited were invented.
In his published ruling, the judge found 31 AI “hallucinations,” “stylistic and reasoning flaws,” and “gibberish” legal analysis. The judge fined the attorneys, then sanctioned the attorneys. Their law firms then fired those attorneys.
Federal judges have established an open-source website where attorneys and judges can cite where AI has been used so that judges, litigants and defense attorneys can be warned of bogus legal work at court. More than 1,300 such AI incidents have been reported so far.
You do not want to be on this list. Nor does your employer.
Yet lots of employers like you are lured by AI’s siren song of speed and efficiency. Instead, you should heed the warnings from courts and federal agencies.
Example #2: In April 2026, the FDA issued a warning letter to a manufacturer. The FDA penalized the company for its procedures, its use of AI to create drug product specifications, and its failure to review those AI-generated documents to ensure they were accurate and compliant.
Example #3: The Big Four consulting firms have all been skewered for recommending AI. I love irony. But no one likes an expert consultant lecturing clients re AI when their reports used “AI hallucinations” and many AI-generated errors of fact and sources. Investigations have had the Big Four retract their reports on the use of AI. You will not survive that investigation by a litigator.
AI’s fingerprints and DNA will be discovered by litigators every time. Litigators will ask the judge to declare your plans, etc., as hallucinations and flawed gibberish because of your use of AI. You have just lost your case. Get out your checkbook.
If you’re in the security and safety world, using AI can very well get you fired and even bankrupt your organization.
Instead, follow recognized standards for creating security and safety plans, then train and exercise those plans. Stay compliant and effective in the old-fashioned way.
Earn it.
About the Author
Bo Mitchell Bo Mitchell
President of 911 Consulting
Bo Mitchell is the President of 911 Consulting. He holds the following designated certifications: CEM, CPP, CHS-V, CBCP, CSI-ML, HSEEP, CSSAS, CNTA, IAC, MOAB, CHSP, CHEP, CSHM, CESCO, CHCM, CFC, CSSM, CSC, CAS, TFCT3, CERT, CHSEMR, CMC
Bo was a Police Commissioner of Wilton, CT for 16 years. He retired to found 911 Consulting, which creates emergency, disaster recovery, business continuity, crisis communications and pandemic plans, and training and exercises for organizations like GE HQ, Hyatt HQ, H&R Block HQ, MasterCard HQ, four colleges and universities, and 29 secondary schools. He serves clients headquartered from Boston to Los Angeles working in their facilities from London to San Francisco. Bo also serves as an expert in landmark court cases nationally.
