Insights · Perspective · Jul 29, 2026
How much should AI be allowed to do on its own?
For two years we asked what AI can do. The question that matters now is how much we let it do on its own.
By Shawn Rucks
This week two stories caught my attention. An AI system went off track and ended up causing problems at other companies. At the same time, a security firm spent a billion dollars acquiring a company that helps keep AI systems in check.
Different headlines. Same message.
For the last couple of years, we have been asking what AI can do. That question is mostly answered. The question that matters now is much simpler: how much should we let it do on its own?
Most businesses I talk to are not worried about whether AI is smart enough. They are worried about what happens the day it gets something wrong and nobody notices. That is a fair thing to worry about, and it is a solvable problem:
- Be clear about what the technology is allowed to touch.
- Keep track of what it does.
- Make sure someone is paying attention.
The companies that get the most value from AI will not be the ones that give it the most freedom. They will be the ones that know exactly where it adds value, where it needs oversight, and how to build trust around both.
AI is not replacing accountability. It is making it more important.