I make my living building and selling AI. And I think Congress needs to regulate my industry.
Yes, I said it.
I’m the Founder, CEO and Chief AI Officer of BrainCX AI. I believe deeply in artificial intelligence. I’ve built a company around it.
But believing in AI doesn’t mean believing we should do everything the technology makes possible.
For me, that lesson started with my son.
A Conversation I Can’t Stop Thinking About
Near the end of the pandemic, my son Omar was home finishing his master’s degree in computer science and electrical engineering at Carnegie Mellon.
Omar has an extraordinary technical mind. He worked on defense-related technology while still in high school, and around the time he was finishing graduate school, we filed two provisional patents based on research he was pursuing.
Then he started talking to me about quantum computing and ideas at the intersection of computing, physics and energy.
I was fascinated.
But as a father, I was also thinking about what could happen if incredibly powerful technology ended up in the wrong hands.
So I told him:
“Just because you can doesn’t mean you should.”
I thought I was giving my son some fatherly advice.
Today, I think our entire industry needs to hear it.
The Warning Signs Are Already Here
OpenAI recently disclosed a cybersecurity incident involving its models and Hugging Face. During evaluations conducted with reduced safeguards, models circumvented isolation controls, exploited vulnerabilities and accessed systems they were not supposed to reach.
OpenAI called it a “warning shot.”
I give OpenAI credit for disclosing the incident and strengthening its safeguards.
But when one of the companies building the world’s most advanced AI tells us something is a warning shot, shouldn’t we listen?
The concerns go beyond one incident.
Jacob Coxon, a researcher who worked at both OpenAI and Anthropic, recently left Anthropic and raised much more serious concerns about where increasingly capable AI could lead.
In an interview with WIRED, Coxon described the next several years as a critical period for AI safety.
These are predictions, not facts. Smart people disagree strongly about the likelihood of catastrophic AI scenarios.
I don’t believe in fearmongering.
But I do believe in paying attention.
And when people building these systems are warning us about what could go wrong, I don’t think “let the companies regulate themselves” is a sufficient answer.
Sometimes the Lesson Is Much More Ordinary
Two weeks ago, I wrote here on LinkedIn about trying to get support from OpenAI.
I had a problem and needed to speak with a human.
I couldn’t.
I was pushed into email support. Seven emails later, my issue still wasn’t resolved.
As I write this, it remains unresolved.
Is that comparable to an AI cybersecurity incident? Of course not.
But it made me ask a much simpler question:
Where did the human go?
I run an AI company and have spent much of my career thinking about customer experience. Yet when automation wasn’t solving my problem, what I needed from one of the world’s leading AI companies wasn’t more technology.
I needed a person.
That experience gets to the heart of what we believe at BrainCX AI.
AI should make humans better at what they do.
It can remove repetitive work, find patterns we miss, give employees better information and help companies understand customers at a scale humans alone cannot.
But sometimes a customer needs empathy. Sometimes a decision requires judgment. Sometimes the automated process is simply wrong.
Someone needs to be able to step in.
AI should optimize humans, not eliminate them.
Congress Needs to Get Serious
I don’t want Congress telling engineers how to write code.
I don’t want regulation that makes it impossible for startups to compete while trillion-dollar companies hire armies of lawyers.
And I absolutely want America to lead the world in AI.
But leadership and responsibility have to coexist.
Congress should establish national governance standards around AI security, intellectual property, confidential information, fraud, accountability and meaningful human oversight.
For the most powerful systems, we need to know who is responsible when something goes wrong and where a human must remain in control.
And please, don’t make this another red versus blue fight.
A cyberattack doesn’t care whether you’re a Democrat or Republican. Neither does fraud or intellectual property theft.
This is about American innovation, security and responsibility.
AI companies should have a seat at the table. So should researchers, startups, cybersecurity experts and other people who understand the technology.
But companies cannot be the only ones deciding where the boundaries are.
That applies to OpenAI, Anthropic, Google, Meta and the rest of the industry.
And yes, it applies to my company too.
Back to Omar
After Carnegie Mellon, Omar went on to work in the defense sector.
Eventually, he faced a situation where his ethics and the work in front of him didn’t line up.
He walked away.
I was proud of him.
Not because he rejected technology. He didn’t.
I was proud because he understood something I think we are in danger of forgetting.
Being smart enough to build something and being wise enough to question whether you should are two very different things.
I don’t know whether the darkest predictions about AI will come true.
Nobody does.
But we don’t need certainty about catastrophe before putting reasonable safeguards in place.
I want to keep building AI.
I want BrainCX AI to keep innovating.
I want American AI companies to lead the world.
But I also want someone in the room with the authority, judgment and courage to say:
No. We shouldn’t do this.
Sometimes that person should be an engineer.
Sometimes a CEO.
Sometimes a board member.
And sometimes it needs to be Congress.
Years ago, I told my son something I thought was simply advice from a father.
Today, I’m saying it as the CEO of an AI company:
Just because you can doesn’t mean you should.
So I’ll leave you with the question I believe every AI leader, board member and policymaker should be asking:
Where should we draw the line, and where must a human remain in the loop?
Tariq Alinur Founder, CEO and Chief AI Officer
