I believe this distinction may have been made by Hayek or some other Austrian.
Essentially information is any data that might influence an act (or specifically the perceived utility or cost of a choice), and knowledge is the information that the actor has already collected prior to the time they made their decision and/or engaged in their behavior.
This helps to explain seemingly irrational behavior that is merely a manifestation of imperfect or asymmetric information (though it doesn't necessarily preclude or eliminate irrationality by any means).
There is also some room for "false" (or conflicting) knowledge, to provide for further irrationality (in the sense that $A$ knows $U(x)>U(y)$ and $B$ knows $U(y)>U(x)$ where $U(x)$ is the average perceived utility of $x$ for all actors in the economy.
He of course does a much better job explaining it than I do.
Econ Journal Watch actually had a symposium for this very question back in 2005. It looks like there are no universally agreed upon definitions of "information" and "knowledge". Nonetheless here's a selection from the three articles in that symposium that I found the most helpful.
- Thomas Mayer
Information [is an] isolated nugget. Knowledge [is] such nuggets of information integrated into coherent constellations, such as generalizations. Understanding is the integration of knowledge into the larger web of our other beliefs. Wisdom goes beyond understanding by adding epistemological and perhaps value and metaphysical judgments, which may, or may not, be articulated.
Mayer gives an example of his distinction between information, knowledge, and wisdom:
The observation that unemployment is at 6% provides information.
The comprehension that this figure has to be interpreted in the light of changes in the number of discouraged workers, the labor force participation rate, the NAIRU, etc., provides knowledge.
Whether this unemployment rate justifies an expansionary fiscal policy requires consideration of vaguer issues (e.g. measurement errors, policy lag). It also requires value judgments. That is, it requires wisdom.
- Israel Kirzner
For Kirzner, the distinction is between "knowledge-as-information and action-knowledge--the latter referring to the knowledge which actually spurs and shapes action."
Alertness is necessary and sufficient for turning information-knowledge into action-knowledge.
Example: Suppose apples are selling for \$0.40 at the west end of town and for \$1 at the east end of town. John and Mary are both given this piece of knowledge-information. However, Mary is not alert and so she does not act on this piece of information. In contrast, John is alert: he realizes he can profit from this arbitrage opportunity and so acts to transform this knowledge-information into action-knowledge.
- Leland Yeager
Yeager gives the example of voters who
are constantly exposed, even without effort on their part, to all sorts of facts and figures and arguments pertaining to policies, proposals, and candidates. Yes; but miscellaneous bits of information, casually acquired and misacquired, do not add up to knowledge, or, better expressed, surely do not add up to understanding.
Interested in how the vote count is going in California, I might ask for the latest information or, more probably, for the latest numbers, but hardly for the latest knowledge.