The risk is "materially understated", researchers are saying as passwords and critical data can be exfiltrated.
When (and why) does AI coding flip from promising to a security nightmare? Let's look under the coding hood.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results