Who Taught the Machine? - Baltimore Jewish Times

Who Taught the Machine?

As artificial intelligence becomes increasingly embedded in the fabric of society, it's not enough to simply marvel at its capabilities. We must grapple with a more fundamental question: who taught the machine? The answers we find reveal deep flaws in our systems of accountability and oversight.

The real danger is not that machines will become sentient, but that humans will lose the ability to understand and control them.

This question isn't academic—it's urgent. From hiring algorithms that discriminate against minorities to predictive policing systems that reinforce racial biases, AI systems are being trained on data that reflects the prejudices of their creators. And those creators? They're often people who never questioned the ethical implications of their work.

The Hidden Cost of Innovation

In my years as an investigative reporter, I've seen how technology companies operate with little oversight, driven by profits and competitive pressure rather than principles. The same forces that drive Silicon Valley's relentless pursuit of innovation often blind engineers to the social consequences of their creations.

We're not talking about rogue algorithms or malicious intent here—this is systemic. When a major tech company deploys an AI system designed to identify potential security threats, but the training data was sourced from predominantly Black and Brown neighborhoods, the result isn't just inaccurate—it's harmful. And when that harm is hidden in complex code and layered with technical jargon, it becomes invisible to most stakeholders.

Who Is Responsible?

Let's be clear: this isn't just about ethics departments or corporate policies. It's about power—about who gets to decide what data gets collected, how that data is used, and what conclusions are drawn from it. When a city contracts with a private firm to deploy facial recognition technology in public spaces, they're not just buying software—they're purchasing access to surveillance systems that may violate the constitutional rights of millions of residents.

And yet, there's minimal public oversight or transparency. The algorithms used are often proprietary secrets, shielded from scrutiny under trade secret laws. What we do know is that these systems have been shown to be significantly less accurate when identifying people with darker skin tones. That discrepancy isn't a bug—it's a feature of a system trained on biased data.

When Accountability Vanishes

One of the most chilling aspects of modern AI development is the way it shifts responsibility away from those who create these systems and onto the individuals who use them. If an algorithm makes a mistake, it's said to be "unfair" or "biased," but never questioned for its fundamental design flaws.

We've seen this pattern play out repeatedly: a hiring platform rejects qualified candidates because of gender or racial bias encoded in its training data; a healthcare AI misdiagnoses patients based on incomplete datasets that exclude certain demographics; a criminal justice system uses risk assessment tools that unfairly target low-income communities. Each time, we're told the machine is neutral and objective—but the truth is far more troubling.

As citizens, we have a right to know how our lives are being influenced by these systems, especially when they make decisions that affect access to jobs, housing, healthcare, and even freedom itself. But too often, those in power hide behind claims of "neutrality" and "objectivity" while failing to acknowledge the role of human judgment and bias in shaping artificial intelligence.

What Needs to Change

Change starts with transparency. We need legislation requiring public disclosure of how AI systems are trained, what data is used, and who has access to that information. We also need independent audits and oversight boards composed of diverse voices—including those directly impacted by these technologies.

More importantly, we must demand accountability from the companies developing these systems. They cannot claim to be neutral when their products systematically harm certain groups. Their responsibility extends beyond profit margins to include the protection of human dignity and civil rights.

This is not about stifling innovation or halting progress—it's about ensuring that progress serves humanity rather than replacing it. When we hand over decision-making power to machines, we must be prepared to accept both the benefits and the costs. But so far, we've been unwilling to pay the full price.

The Price of Ignorance

In Baltimore, where I've spent considerable time covering issues related to social justice and criminal reform, the story of AI and its impact on communities of color has become increasingly urgent. These aren't theoretical concerns—they're lived realities that shape people's lives every day.

As we move forward into an age where algorithms will make more and more decisions affecting our daily lives, we must ask ourselves: are we creating a future where technology serves the common good or one where it reinforces existing inequalities? The stakes couldn't be higher, and the clock is ticking.

We cannot allow another generation to pass without taking meaningful action to ensure that artificial intelligence does not become another tool of oppression. We owe it to ourselves, to our children, and to future generations to demand better.

• AI systems must be transparent and auditable
• Data used for training should reflect diversity and accuracy
• Corporations must face legal consequences for discriminatory practices
• Public oversight is essential for democratic accountability

These are not impossible goals—they're necessary ones. If we fail to act now, the machine will have taught itself well enough to ensure that it never has to be accountable again.