Observation·
AI is starting to act without being asked
Until now, AI did what you asked. You typed a question, and it answered. You gave it a task, and it did the task.
That is how we used AI at work for the last four years. My teams used AI agents to write and test software. The agents did the work. But a person always told them what to do next.
That is changing. New AI agents do not wait to be asked. They search for options, compare them and negotiate for you. Often, the other side of the deal is an agent too.
We have seen shifts like this before. The internet started as something new, and became something everyone uses without thinking. Smartphones did the same. After five years, AI is becoming normal in the same way. The next step is AI that acts for you.
That raises a simple problem. When two agents agree on a deal, no person was there. But the deal still needs a person who is responsible for it. An agent can make a decision. Nobody can hold an agent responsible.
So the question has changed. It used to be what AI can do for us. Now it is who is responsible for what AI does in our name.
Reel transcript · 0:57
Until now, AI did what you asked. You typed, and it answered. That's about to change.
You ask. AI answers.
For four years, my teams used AI agents to build software. The agents did the work, but a person always told them what to do next.
AI acts. For you.
Now agents don't wait to be asked. They search, compare and negotiate for you. And often, the other side is an agent too.
Two agents agree. Who is responsible?
We've seen shifts like this before: the internet, then smartphones. Each one started new, and became normal. AI is becoming normal now. Next, it acts for you.
When two agents agree on a deal, no person was there. But someone still has to be responsible for it.
The takeaway
So the question has changed. It's no longer what AI can do for us. It's who is responsible for what it does in our name.