Most people currently use AI in a fairly simple way.
We ask questions.
We generate ideas.
We improve writing.
We summarise information.
The final decision still belongs to us.
But that may be starting to change.
Recently, Visa announced that it is working towards allowing AI assistants such as ChatGPT and similar systems to make purchases directly using a customer’s credit card.
At first, that might not sound like a huge development.
Until you stop and think about what it actually means.
Today, if you ask AI to recommend a product, it might give you a list of options.
You still decide what to buy.
You still click the button.
You still enter your payment details.
The AI provides information.
You take the action.
What Visa is exploring could move AI into a very different role.
Instead of simply recommending products, AI could eventually:
- compare prices
- find deals
- select products
- place orders
- complete purchases
All on behalf of the user.
In the technology world, this is often referred to as “agentic AI.”
The idea is that AI doesn’t just answer questions.
It performs tasks.
That may sound convenient.
And in many situations, it probably would be.
Imagine asking:
“Find me the cheapest flight to Madrid next month.”
Or:
“Order my usual groceries.”
Or:
“Replace my printer ink when it runs low.”
An AI assistant could potentially handle much of the process automatically.
But the announcement also raises some interesting questions.
How much decision-making are people comfortable handing over?
Would you trust AI to choose a product for you?
Would you trust it to spend your money?
What happens if it misunderstands your instructions?
Or purchases something you didn’t intend to buy?
These aren’t technical questions.
They’re human questions.
Because the challenge isn’t simply whether AI can complete a transaction.
The challenge is trust.
Most people are comfortable asking AI for information.
Allowing AI to access a credit card feels very different.
In many ways, this could represent the next stage of AI adoption.
The first phase was information.
The second phase was assistance.
The third phase may be action.
And that’s where things become particularly interesting.
Because once AI starts acting on our behalf rather than simply advising us, our relationship with the technology changes.
It becomes less like a search engine.
And more like a digital assistant with real responsibilities.
Whether people embrace that idea remains to be seen.
But one thing seems clear.
The conversation around AI is slowly moving beyond what AI can say.
And towards what AI can do.
Curious — would you trust an AI assistant to make purchases using your credit card if it saved you time and money?
No comments:
Post a Comment