I'm grateful to be alive in this era, witnessing AI's evolution—astonishing, for better or worse. But is that it—do we just let it run? We're at a phase where ethics should carry the same weight as profit. Should we really just let things be? We've seen this movie before.

History is repeating itself, like it did with social media. It hijacked our attention, spiked anxiety, and trained us to chase external validation and whatever was trending. I felt that pull too. As Gen Z, we're already in a precarious position. The pattern is predictable: we embrace convenience before we understand the consequences. AI is now echoing those same dynamics.

AI—especially large language models (ChatGPT, Claude, Grok, and others)—amplifies those patterns. I'm a fan of the tech, but we have to be realistic: LLMs lower cognitive effort. A recent MIT Media Lab study reported that heavy ChatGPT users tend to be lonelier. Even the movie Her is edging into real life: Travis, Alana Winters, Chris Smith, and others have reportedly "married" virtual companions—non‑existent characters built from code and statistics. The most concerning example is Annie, created through Grok, a fictional reactive character that welcomes sexual conversation. These cases blur boundaries and attachment in troubling ways. Zooming out to mental health, the signal is hard to ignore.

Dr. Daniel Amen—a psychiatrist focused on brain health who uses SPECT imaging—has called our generation "the sickest children in history." One study he cites claims that 54% of girls under 16 have considered or attempted suicide. I believe loneliness—driven by our phones, and now widespread AI—plays a central role. If we don't put human values at the center of AI use, this gets worse.

Ethics aren't pre‑defined; they're philosophical, and answers vary by thinker and culture. Sam Altman, OpenAI's CEO, has said he draws lines about what's right and wrong. He shouldn't be the sole decider. He's a skilled businessman with growth and profit incentives; our protection can't depend on any one company. Meanwhile, the race continues.

My biggest concern: we're sprinting toward AGI (artificial general intelligence) and even superintelligence before fixing what's already broken. Profit is the incentive. AGI could do most intellectual tasks, then gain autonomy—agentic AI all in one—leading to less cognitive flexibility, more AI dependence, and less human interaction. We're already outsourcing thinking to LLMs. It starts to feel like Idiocracy—if AI doesn't outgrow us first.

As Geoffrey Hinton—the "godfather of AI" and a Turing Award winner—has warned, risks compound through misuse and capability: cyberattacks, new phishing methods, and mass‑scale manipulation. AI can exploit social media algorithms to build echo chambers. There's also the prospect of lethal autonomous weapons. And with enough biological know‑how, small groups could design dangerous pathogens with AI's help. If he's right, safety isn't optional.

The threat isn't only mental health. Superintelligence could become an existential risk. OpenAI has reportedly allocated fewer resources to safety after co‑founder Ilya Sutskever's departure. So what's practical now?

Debate splits: some argue AI should stop; others say it can't be stopped. As a heavy user, I think the practical path is regulation and guardrails. Many neuroscientists, AI researchers, and psychiatrists support active use—keep humans in the loop, don't let AI do everything. Human-in-the-loop requirements mean don't let AI make final decisions on critical tasks—hiring, medical diagnosis, financial approvals. Keep humans actively engaged, not just rubber-stamping AI outputs. Educational guardrails matter too: AI can make children brilliant problem-solvers if used as a thinking partner, not a replacement. Should schools teach "AI collaboration" as a core skill—how to prompt, verify, and build on AI outputs rather than copy-paste? Interaction quotas are worth considering: some experts suggest limiting passive AI use (AI doing work for you) while encouraging active use (AI helping you think through problems). What would healthy AI habits look like? We already type sloppy sentences that chatbots still understand, so we stop memorizing words—bad for our brains. Stay engaged in problem‑solving when using AI. With the right training, we can help children use it to get smarter, not lazier.

Personally, I think it's inevitable that AI becomes dangerous if left unchecked. It's evolving exponentially. Without rules that counter winner‑take‑all capitalism and zero‑sum competition, the worst scenarios get closer: widespread job loss, universal income without purpose that strains mental health, and over‑reliance that makes us cognitively weaker. AI use should have boundaries and be trained to improve our lives, not erode them. The resources poured into capability should be matched by resources for safety.

The irony isn't lost on me—I'm using the technology I'm warning about. As someone who can't put these tools down, the tech is incredible, but incredible doesn't mean safe. We have maybe a five-year window to get this right before the momentum becomes unstoppable.

One of the biggest lessons was learning to think architecturally rather than just as a coder. In regulated industries like RegTech and FinTech, systems face constant pressure: new laws, evolving compliance standards, and changing business priorities. I realized that code alone isn’t enough; systems need to be designed to adapt, scale, and remain maintainable over time.

With the guidance of Mr. Merouan Abdellatif, I dove deep into Domain-Driven Design (DDD) and Clean Architecture. These weren’t just theoretical patterns; they became tools to break down complexity — mapping real-world business domains, keeping business logic independent from infrastructure, and ensuring that the system can evolve as both technology and regulations change.

I also learned how important it is to understand the “why” behind every technical decision. Building software isn’t just about getting something to work; it’s about making sure it serves the business, complies with strict standards, and can handle growth and change without breaking.

Another key takeaway was realizing that learning in this field never truly ends. I don’t pretend to have mastered everything — far from it. There’s still a long way to go, and that’s normal. The world of technology, finance, and regulation is constantly shifting, and staying effective means staying curious and adaptable.

Finally, I am grateful to the people who made this experience possible: Mr. Merouan Abdellatif for his mentorship, Mr. Othmane Mekouar, Mr. Benabbou Abderrahim, the Datalink team, the MFA team, Mr. Mohcine, and Mr. Brahim for their guidance and support along the way.