All editions

No. 1Technology & AI2 min read

What If the AI Gets It Wrong?

Translated from the Spanish original.

Leer en españolOriginally published on LinkedIn (opens in a new tab)

What happens when a bank denies a loan because of a biased algorithm, when an HR system screens out candidates with "non-standard" profiles, or when facial recognition mistakes an innocent person for a suspect?

What level of responsibility does the AI bear?

This question matters more than ever, because most legal frameworks rest on a simple premise: a person decides, a person answers for it.

But that premise falls short in a world where the decision is made by a system, trained by others, on data nobody fully understands… and that is where the law starts to stutter.

What do different jurisdictions say?

· European Union: the recent AI Act bans certain high-risk uses and imposes rules on developers; in other words, it limits how far these systems can manage and govern themselves.

· United States: the debate centers on whether the creator of an algorithm should be liable like the manufacturer of a defective product.

· Latin America: the rules are fragmented, but the errors are already happening… and the lawsuits are just around the corner.

Why does this matter?

Because many professionals, companies and governments are delegating decisions to artificial intelligence models, or relying on them, and in many cases they do not fully understand the risks involved.

· Lack of transparency (why was my loan rejected?)

· Biases that can trigger discrimination claims

· Difficulty proving fault or correcting the error

THE FINE PRINT

Many contracts already include clauses on the use of automated tools. They may look like simple clauses, but their consequences are anything but automated: lawsuits, fines and even reputational crises.

While technology moves at the speed of innovation, the law moves with feet of lead. But in between, someone always pays the bill. And so far, no AI has ever had to hire a lawyer. However wonderful and helpful a tool it is, it is not free of errors, which is why the human factor remains essential: AI can help you, but it cannot replace you. At least, not yet.

Did you enjoy this edition?

The Fine Print does not try to predict the legal future… but it does help you read it better.

Share your thoughts with me.

Questions about your case? Book a free 30-minute call

Hassim reviews your situation and tells you which route fits, in English or Spanish.

Book a free call