On-site interviews are pretty clearly the antidote to this, but that costs money this quarter. In contrast, the cost of hiring a total fraud cannot be pinpointed to any one metric or line item.
Good interviews are the antidote to this. Outside of certain fields ability to do leet code puzzles does not correlate with job performance. An interview where you do some realistic work is the direction we should be moving. Build some small functionality that’s similar to what the company you’re interviewing for does. Let them use tools they enjoy, ai or not. If they fly through it expand the tasks. Talk about why the code they, or an ai wrote works. The tradeoffs in their designs. I’ve given hundreds of interviews without asking a single leet code style question and not regretted it.
More importantly it can be blamed on the fraudster, and acountability for who missed the fraud is no longer traceable.
I’m not sure why it matters, just give people problems to solve and let them use whatever tools are available