Suvianna Grecu: Strong Rules Are the Only Way to Prevent an AI Trust Meltdown

 

Racing Into the AI Age: Why Speed Without Safety Could Cost Us Our Trust in Technology

Around the world, nations, companies, and innovators are locked in a race to build and deploy artificial intelligence. Every month seems to bring a new breakthrough—a chatbot that talks like a person, a diagnostic system that can detect diseases earlier than any doctor, an algorithm that can write, design, and even compose music.

But amid this rapid acceleration, one of the most respected voices in AI ethics is sounding the alarm: if we prioritize speed over safety, we may be steering ourselves straight into what she calls a “trust crisis.”

---

The Warning from Suvianna Grecu

Suvianna Grecu, founder of the AI for Change Foundation, has spent years at the intersection of technology, governance, and social impact. She’s worked with governments, advised companies, and spoken at global conferences about how AI can serve humanity rather than harm it.

Her message today is sharper than ever:

“Without immediate and strong governance, we are on a path to automating harm at scale.”

This is not about fearing the technology itself, she clarifies. AI is not inherently good or bad—it’s a tool. The danger comes from how we deploy it and how little structure exists around its rollout, especially when these systems start making high-stakes decisions.

---

The Silent Spread of AI into Critical Systems

Grecu points out that AI has quietly moved from powering harmless recommendations—like what movie to watch next—to making life-altering decisions in critical sectors.

It can now determine:

• Whether someone gets a job interview

• Who qualifies for a loan or mortgage

• How a patient’s symptoms are prioritized in healthcare

• Who is flagged for further investigation in criminal justice systems

The problem? Many of these systems are rolled out without adequate testing for bias, without a clear understanding of their long-term impact, and without enough oversight to catch mistakes before they cause damage.

Bias in AI is not just theoretical. There have been real-world cases where automated hiring tools downgraded applicants with certain names, credit scoring algorithms penalized minorities, and predictive policing tools disproportionately targeted specific communities.

When decisions are automated at scale, even small errors or hidden biases can harm thousands—or millions—of people before anyone notices.

---

The Ethics Gap: From Principles to Practice

Grecu believes the biggest ethical danger is not that AI is too advanced, but that our ethical structures are too weak. Many organizations have “AI ethics principles” written into glossy reports. These often include admirable commitments to fairness, transparency, and accountability.

But when you look closer, these principles rarely translate into daily operational practices. In other words, they exist on paper, not in the actual workflows where AI is designed, trained, and deployed.

This “ethics gap” is where the real risk lies. Without clear accountability—without a named person or team responsible for outcomes—good intentions can dissolve into empty promises.

---

From Abstract to Action: The AI for Change Approach

Grecu’s foundation, AI for Change, is on a mission to close that ethics gap. Instead of treating AI ethics as a philosophical debate reserved for think tanks and academic journals, she wants to turn it into a set of practical, repeatable tasks—as normal and necessary as quality control in manufacturing or safety checks in aviation.

Some of the tools she advocates include:

• Ethics Checklists: Built into the design process so that every developer must consider potential harms, biases, and societal impact before writing a single line of code.

• Mandatory Pre-Deployment Risk Assessments: Just as you wouldn’t release a new drug without clinical trials, no AI system should go live without stress-testing for bias, reliability, and fairness.

• Cross-Functional Review Boards: Bringing legal, technical, and policy experts into the same room to evaluate AI projects from multiple perspectives.

• Clear Ownership: Assigning responsibility for AI decisions to specific individuals, making accountability traceable.

This isn’t just about protecting people from harm—it’s also about building public trust, which Grecu sees as the most valuable currency for technology companies in the coming decade.

---

Why Trust is the Real Battleground

Trust is fragile. Once lost, it’s almost impossible to rebuild.

Grecu warns that if AI systems repeatedly produce unfair or harmful outcomes—especially in areas like healthcare, finance, and law—public confidence in the entire technology could collapse. This “trust crisis” wouldn’t just hurt the companies involved; it could slow adoption of AI innovations that might otherwise have saved lives, improved efficiency, or solved critical problems.

We’ve seen similar dynamics play out before:

• Social media was once celebrated as a tool for global connection, but after years of privacy scandals, disinformation, and political manipulation, public trust has eroded.

• Self-driving cars made headlines for their promise of safety, but high-profile accidents have made many people wary of stepping into one.

If AI suffers the same fate, the damage could be far greater—because unlike a social network or a single product, AI is becoming embedded in the core infrastructure of society.

---

Governance: Not Just Government’s Job

When it comes to enforcement, Grecu is adamant: responsibility cannot fall entirely on one side.

“It’s not either-or, it has to be both,” she says of the roles of government and industry.

Governments must set the legal boundaries and minimum standards, especially where fundamental human rights are involved. Regulation is the floor, the basic safety net that ensures no AI system drops below a certain standard of fairness, transparency, and safety.

However, the private sector has a crucial role too. Companies have the technical expertise, agility, and innovation capacity to go beyond compliance. They can:

• Develop advanced auditing tools that continuously check for bias or drift in AI models.

• Create user interfaces that clearly explain AI decisions in plain language.

• Experiment with new methods of safeguarding against manipulation or misuse.

Leaving governance only to regulators risks slowing down innovation. But leaving it only to companies risks abuse, shortcuts, and unchecked profit motives. The only sustainable way forward is collaboration.

---

The Long-Term Risks We’re Not Talking About Enough

Grecu’s concerns go beyond today’s headlines about AI bias or job automation. She is deeply worried about emotional manipulation—the ability of AI to influence human thoughts, feelings, and behavior in ways we may not even notice.

Already, algorithms shape what news we see, what products we buy, and even what political messages we’re exposed to. As AI systems become more sophisticated at reading emotional cues and tailoring responses, this influence could become far more powerful—and far harder to resist.

This raises serious questions about personal autonomy. How much of our decision-making will truly be our own if machines can subtly nudge us toward certain choices without us realizing it?

---

Technology is Never Neutral

One of Grecu’s core beliefs is that technology is not neutral. Every AI system reflects the values of its creators—whether intentionally or not.

“AI won’t be driven by values unless we intentionally build them in,” she warns.

It’s a mistake to think AI simply mirrors “the world as it is.” In reality, it mirrors the data we feed it, the objectives we give it, and the outcomes we reward. Without conscious intervention, AI will naturally optimize for what is easiest to measure: efficiency, scale, profit.

But efficiency is not the same as justice. Scale is not the same as dignity. Profit is not the same as democracy.

---

Europe’s Critical Opportunity

For Europe, Grecu sees a unique moment in history. The region has a long tradition of embedding human rights, transparency, sustainability, inclusion, and fairness into policy. If these values are built into AI from policy to design to deployment, Europe could set a global standard.

This is not about halting progress or slowing innovation. It’s about ensuring that as AI becomes more powerful, it remains aligned with human well-being—not just market success.

---

Shaping the Narrative Before It Shapes Us

Grecu ends with a challenge:

“We need to take control of the narrative and actively shape it before it shapes us.”

The AI race will continue—there’s no slowing the momentum now. But whether we cross the finish line to a better future or a fractured society depends on the guardrails we build today.

The choice is not between progress and safety. The choice is whether we define progress only in terms of speed, or whether we have the courage to measure it in terms of trust, fairness, and shared benefit.