Does AI make IT specialists less needed, or more?

One of America's biggest banks just hired a thousand extra people to clean up after an AI. That is not the headline you expected.

Everyone has heard the prediction. AI writes the code, AI finds the bugs, so the humans in IT quietly fade out. Now look at what actually happened at one very large bank.

Its chief executive explained that in the second quarter the bank brought in about 1,000 contractors to patch software flaws uncovered by advanced AI models, Anthropic's Mythos among them. They work alongside roughly 3,000 people the bank already had on this job, and they were still there months later. Together with a big anti-money-laundering bill, the cleanup cost a couple of hundred million dollars in a single quarter. One scan alone reportedly flagged more than 23,000 issues.

Finding is cheap. Fixing is not.

That is the whole story in a sentence. AI is astonishing at spotting problems. It does not make the work of fixing them disappear, because a fix is more than a line of code.

Hold on. If AI found the bugs, why can't AI just fix them? Why hire a thousand people?

Because a patch can be written in an afternoon and still take a week to ship safely. In a bank every change needs regression tests, a slot in a maintenance window, a conversation with the vendor who supplied the software, and someone accountable if it breaks the payments system. That is judgement and coordination, and it lives in people.

Regulators are adding to the load. One central bank has given banks until the end of October to produce plans for defending against AI-driven cyber threats.

So here is the honest picture. The demand for IT people is not shrinking, it is changing shape. Fewer people to find the needle, many more to triage, test, roll out and take responsibility. Different skills, same hunger. And if every company that has not yet run the scan runs it next year, that hunger gets bigger, not smaller.