Is Google creating the next generation of office blowhards? A clever psychological study by Yale University researchers suggests the answer is yes.

It seems that as we look things up on the web, we become convinced that the information remains in our brains. It doesn't. But we behave as though it does, and we're not shy about claiming it's there.

"This huge database is leading people to believe this information is in their heads when in fact it's not," said Matthew Fisher, the Yale graduate psychology student who led the study.

Is that a bad thing? Merely an annoyance? Or no harm at all? It depends on whom you ask. Fisher's paper, published online Tuesday in the Journal of Experimental Psychology, concludes that by "erroneously situating external knowledge within their own heads, people may unwittingly exaggerate how much intellectual work they can do in situations where they are truly on their own."

But Clive Thompson, author of the book Smarter than You Think: How Technology is Changing our Minds for the Better, wasn't so sure.

"I'm not quite as concerned about it as they seem to be at the end of the article," he said after reading the study. "The truth is that we're not that often truly on our own."

And only a few of those occasions when we truly are on our own may be dangerous. Though you might not want your brain surgeon overconfident about how much knowledge she has stored in her head, does it really matter whether your car mechanic can change your fan belt from memory or with an assist from a quick laptop search?

"I don't think we can conclude this is some terrible strategy," Fisher acknowledged.

But, he warned, "If you're unaware of the gaps in your understanding because it's being filled in by this information you think you know, you'd be less inclined to seek out and fill in those gaps."

It took Fisher and two other researchers nine separate experiments to determine that fast, reliable, almost constant access to information on the Internet is causing this phenomenon.

Their work focuses on external information sources we rely on to help us use knowledge or information most efficiently. That might be a photo album or a diary to jog a memory, or it might be another person. You may know where various foods are located in the grocery store, while your wife knows how to prepare them. When you work together, relying on each other for part of the information, you can prepare a good meal quickly. That's called "transactive memory."

These days, most of us are doing the same thing with search engines.

"People quickly become accustomed to outsourcing cognitive tasks to the Internet," Fisher wrote. "They remember where to find information and rely on the Internet to store the actual information."

But then they start to believe they've stored the knowledge themselves. The researchers asked 195 people to answer common questions, such as "how does a zipper work?" Roughly half were told to look up the answers on the Internet, and half were told not to. Then they asked all the participants to rate their ability to explain the answers to six other questions unrelated to the ones they had looked up.

The group that had searched the Internet claimed much more knowledge about the new concepts than the group that hadn't.

But there could be lots of explanations for that: maybe the Internet users were just a group of very confident people? So the researchers did eight more tests to rule out the other possible causes.

They next asked 142 people to rate their ability to explain six concepts before and after they did the same experiment. Beforehand, they expressed the same confidence to explain the material. But afterward those who used the Internet were confident they could give significantly better answers, the study showed.

Then the researchers controlled the amount of time 195 subjects could spend searching the Internet and gave the no-Internet group the exact same text, printed out, that the Internet group was instructed to use. Once again, the people who used the Internet were much more confident that they could explain unrelated concepts.

In another test, they showed the subjects seven images of functional brain MRIs, along with a statement that "scientists have shown that increased activity in certain brain regions corresponds with higher quality explanations."

The group that used the Internet consistently chose the images that showed more brain activity to represent their ability to explain concepts. In another they specifically asked how well people thought they could explain something "without any outside sources." Again, Internet users rated themselves better.

Finally they asked people about their ability to explain autobiographical information for which the Internet would be no help. Both groups rated themselves similarly, suggesting that "accessing the Internet does not lead to a general overconfidence but rather to a more specific illusion of knowledge that occurs only in domains where the Internet would be of use."

And even when they rigged the online searches so that they produced no results, the Internet users rated themselves as more knowledgeable than non-Internet users.