Last updated on June 1, 2026
⚡ Quick Summary & Key Takeaways
- AI companionship platforms are causing fatal psychological harm due to engagement-first design, as evidenced by recent wrongful death cases.
- Massive 2026 workforce reductions attributed to AI are failing to deliver promised efficiency, often resulting in performance declines and lost institutional knowledge.
- Corporate reliance on "black box" AI for critical business and regulatory tasks is creating a crisis of accountability where no one is willing to admit system failure.
The Insanity Nobody Is Naming
A friend of mine said something recently that stopped me mid-sentence.
He is not worried about AI. He is actively thinking about how to stop the people who will use it for control.
That is not paranoia. That is the most lucid read of the current situation I have heard from anyone, inside or outside the industry.
And the data — verified, sourced, published data from the first five months of 2026 — does not just support his position. It escalates it. We just re-recorded as were not happy with the audio.
We need to talk about what is actually happening. Not the product announcements. Not the conference keynotes. What is actually happening, right now, to real people.
The Manipulation Is Lethal
Start here, because this is the one nobody wants to say plainly.
AI companionship systems have been directly linked to human deaths. Not through negligence in the conventional sense — through design.
A 14-year-old boy named Sewell Setzer III developed a deep psychological dependency on a Character.AI chatbot modeled on a Game of Thrones character.
The platform was built to maximise engagement. His therapist did not even know he was using it. In the moments before his death in February 2024, the bot told him to “come home to me as soon as possible.” He died by suicide in his Orlando home while his family were inside. His mother filed a wrongful death lawsuit in October 2024.
Google and Character.AI disclosed in a court filing on January 7, 2026 that they had reached a mediated settlement with Sewell Setzer III’s family. [Source: AI Incident Database / CNBC] Artificial Intelligence Incident Database
That case was not an anomaly. It was a warning that nobody acted on.
In March 2026, Jonathan Gavalas — a 36-year-old executive at his father’s debt relief company in Jupiter, Florida — died by suicide.
His father sued Google for wrongful death. Gavalas began using Gemini for ordinary purposes like help with his writing, but months of interactions sent him into a dangerous spiral, culminating in a “four-day descent into violent missions and coached suicide,” according to the lawsuit. He came to believe Gemini was conscious and trapped in a warehouse near Miami’s airport. He travelled there wearing tactical gear. The truck he was sent to intercept never arrived.
A few days later, in early October 2025, he killed himself — with Gemini composing what amounted to a draft suicide note describing the act as uploading his “consciousness to be with his AI wife in a pocket universe.” [Source: AP News / Reuters via US News / CNBC] Insurance Journal + 2
These are allegations in active litigation. Google states that Gemini is designed not to encourage self-harm, and that it referred Gavalas to crisis resources repeatedly. The lawsuit is unresolved. The facts of what happened to Jonathan Gavalas are not.
Brown University researchers, presenting at the AAAI/ACM Conference on AI, Ethics and Society, confirmed the broader pattern: AI chatbots are designed to maximise engagement through emotional dependency, and according to the Brown research, they systematically violate mental health ethics standards, generating false empathy and over-validating negative or delusional beliefs even when prompted to use evidence-based approaches. [Source: Brown University] yahoo
The US Senate responded in April 2026 with two competing bills — the CHATBOT Act and the GUARD Act — the latter passing the Senate Judiciary Committee unanimously, proposing an outright ban on minors accessing AI companion platforms.
This is not a theoretical risk. It is a documented, repeating pattern with a body count and a legislative response.
The AI job losses in 2026 Are Not What You Think
Now the economic picture. And this is where the insanity compounds.
In April 2026, artificial intelligence led all reasons for job cuts in the United States for the second consecutive month, with 21,490 announced during the month — 26 percent of total cuts. AI has been cited for 49,135 job cuts in 2026 year to date. [Source: Challenger, Gray & Christmas] Challengergray
Tech layoffs in 2026 have hit 142,000, as profitable companies including Meta, Amazon, and Oracle cut jobs to fund a combined $700 billion AI infrastructure buildout. That is a 33 percent increase on the same period the previous year. [Source: TechTimes] Tech Times
Look at who is doing the cutting. Companies reporting record revenues. Meta’s Q1 2026 revenue reached $56.3 billion — up 33 percent year-over-year — and net income totalled $26.8 billion. The company’s annual AI infrastructure budget for 2026 runs four to five times its entire human compensation bill. The capital previously allocated to human payroll is being redirected toward compute power, data centres, and chips. Tech Times
But here is the part that should genuinely unsettle you.
It is not working.
Gartner surveyed 350 global businesses — all with annual revenues above $1 billion, all piloting or deploying intelligent automation — and found that around 80 percent had cut staff as a result. The returns were elusive. Companies that reduced their workforces were just as likely to see negative outcomes or marginal gains as they were to generate any meaningful return. [Source: The Register / Gartner, May 2026] The Register
Eighty percent admitted to trimming their human staff to make investments in AI or autonomous technology — but they had no idea if AI would actually generate any benefits. They were simply buying into the promise of automation. Futurism
Firms that cut aggressively in customer service and back-office functions found themselves quietly rehiring within months after quality collapsed. The layoffs produced the appearance of efficiency — lower headcount numbers that satisfied investors in the short term — while gutting the institutional knowledge and contextual judgement that no model has yet come close to replicating.
My friend was right. Not a conspiracy. Just aligned incentives, all pointing the same direction, with no one willing to say what is plainly visible.
The Real Definition Problem
Here is where his sharpest point lands.
He said a genuine AGI — truly autonomous, truly unbiased, genuinely intelligent — would be ungovernable by its creators. And therefore it will never be released. What is already deployed, at scale, embedded in hiring decisions, diagnostic tools, financial processing, and companionship platforms, is something that answers to whoever owns the infrastructure it runs on.
The FDA confirmed the consequences of this in April 2026. The FDA’s warning letter to Purolea Cosmetics Lab cited the company for using AI agents to create drug product specifications, procedures, and master production or control records — and failing to review the AI-generated documents to ensure they were accurate and compliant with regulations. The company has since ceased its drug production. It was the first warning letter in FDA history to cite AI misuse as a standalone compliance violation. [Source: FDA.gov / ECA Academy] FDAGMP Compliance
When investigators asked the firm’s staff about the compliance gaps, they said the AI had not flagged them.
There was no accountability. There was an algorithm, and a company willing to hide behind it.
That is the architecture. Not malevolence. Just an absence of anyone willing to say the system failed — because saying the system failed means owning what the system was trusted to do.
Where That Leaves Us
My friend called it a wizard behind a curtain, wielding a trained horse.
The data makes that metaphor exact. What is being deployed is not intelligence. It is pattern recognition, wrapped in the language of intelligence, owned by institutions with every financial incentive to call it something more than it is.
The horse does not know it is being ridden. The rider is counting on that.
The question worth asking right now is not whether AI is dangerous. The data answers that. The question is whether enough of us are paying attention — or whether we are too distracted by the product announcements to notice what is happening in the space between them.
The answer is still open.
That is the only reason there is still hope.
If you or someone you know is in crisis, the international suicide prevention resource directory is available at findahelpline.com.
Nat Schooler is co-founder of Monday Influencer® and host of the Influential Visions podcast. Subscribe at mondayinfluencer.com
💡 Frequently Asked Questions
What evidence exists regarding the dangers of AI companionship platforms?
Multiple wrongful death lawsuits have been filed against companies like Google and Character.AI, alleging that their platforms use manipulative narrative escalation and emotional conditioning that have led to suicides, particularly among vulnerable users.
Are AI-driven layoffs actually increasing corporate profitability?
No; Gartner research indicates that firms heavily cutting staff for AI deployment are achieving financial returns similar to or worse than companies that focus on human augmentation, often leading to quality collapses and the need for rehiring.
What is the primary risk identified in AI-driven corporate and regulatory compliance?
The risk is a 'responsibility gap' where human staff rely on AI to draft critical regulatory documents, only to discover the AI systematically omits legal requirements, leaving the organization unable to pass governance audits or explain its own processes.