AI is not stealing lawyers' jobs. It is being used to manipulate decisions nobody audited

AI is not stealing lawyers' jobs. It is being used to manipulate decisions nobody audited.

In 2024, a law firm in the US submitted precedents in a federal case. Real cases. Solid arguments. The documents had been altered by AI to favor the client's thesis. The judge only found out because opposing counsel knew the cases by heart.

This is not fiction. It is what the Brazilian judiciary is starting to face now.

And the perimeter of the problem is not legal. It is governance.

If AI manipulates documents in court, it also manipulates what supports the board's table:

• contracts reviewed without cross-checking
• technical reports read by a model without protocol
• due diligence nobody reread at the source
• strategic reports nobody validated at the origin

The executive who outsources critical reading to AI without corporate governance over the tool is not gaining productivity. He is signing a decision he did not audit.

AI governance is not an IT agenda item. It is a board agenda item.

Whoever decides without understanding what the tool does, and what it can be used to do, is not using AI. He is being used by it. The cost of that omission has already started to show up in court records. It will show up in balance sheets.

There is no recommendation. There is a decision.