(FORTUNE Magazine) – An e-mail-savvy executive--worth well over $12 million--relies on a simple techno-trick to keep his people on their toes. The instant someone commits to a deadline, he sends a confirmatory e-mail. Then he spends another 30 seconds composing a quickie follow-up requiring a thoughtful response. He then forgets about it.

The gimmick? He has programmed his laptop to automatically send this follow-up memo three to five days before the deadline. As a classy added touch, the tickler gets postmarked between midnight and 2 A.M. to make it appear as if the executive is working late. He's not; his computer is.

To co-workers he's downright spooky. His timely e-mails make him appear to be on top of every single commitment. A few of his sharper colleagues are suspicious, he acknowledges, but they still can't be sure which e-mails are live and which are Memorex. So they respond. Warily.

Is this little managerial ruse a harmless white lie? Or is it digital deception that breeds cynicism and distrust? The $12 million man doesn't get the question: "What's the big deal? I'm not really misleading anyone. I'm just using the network to be there when I can't. Do you really think I'd be a better manager if I personally sent the e-mail at midnight the week before?"

Technology, like appearances, can be deceiving. What's so counterintuitive, however, is that deception seems to be essential to making technologies user-friendly. Caller ID and voice mail screen out unwanted callers. E-mail minimizes contact with unpleasant colleagues. Bcc:'s--blind carbon copies--alert corporate allies that an e-mail has been sent without letting the intended recipient know who else is in on it.

In the good old days, managers asked their secretaries to lie; today, managers are learning how to ask their technologies to lie; tomorrow, their technologies will advise them on what deceptions yield the best results. It's only a matter of time before Mr. $12 Million's laptop is actively exchanging faux responses over the network about the faux inquiries it sent. But better our machines lying to each other than to us, right?

In the Web world there's tremendous excitement about "bots": techie slang for "knowbots" (think robots for the Internet) that can traverse cyberspace to track down a piece of information, a price, a transaction, a bid, or a schedule. Just as you can easily program your telephone to keep calling a number that's busy, you can program a bot to bid in an eBay auction, sell a stock, or book an appointment. Bots are little software agents that will literally do your bidding.

They're coming soon to an Intranet near you. They promise to make your work life easier and more productive. How will they do it? The truth is that lying is often the best strategy for bots. Trust me, you don't want a bot that tells the truth, the whole truth, and nothing but the truth. Under many circumstances, it makes perfect sense for network bots to generate decoys, phantom tasks, and other absolutely false information--let's call them "fauxbots"--in order to protect their humans. To wit, your meeting scheduler fauxbot might claim there will be eight people at your meeting (when there are really only five) in order to procure that conference room you like so much. Fauxbots can also "lie" to cut ahead in line or take priority in resources. Why? Because not lying actually hurts the client. Sophisticated fauxbots won't even have to be programmed to cheat; they will "learn" what to do by tracking their client's network behavior.

But fauxbots create obvious quandaries for networked organizations. When competing bots duel to obtain scarce resources, assuming honesty is simply naive. Should firms simply outlaw fauxbots? Good luck. Can organizations count on ethics to get their executives' bots to behave? Human nature says no. You have inevitable tensions between bots doing what's best for their "clients" and doing what's best for the organization. Should bots optimize the client or the network good? "This becomes a social dilemma," acknowledges Bernardo Huberman, a Xerox PARC scientist who has built agent-based computer-network simulations. "One general result we've had is that, as the group of [bots] gets larger, these problems of deception emerge. There is a critical mass beyond which it is impossible to sustain cooperation."

In other words, honesty and cooperation break down when too many bots pursue the best strategy for their individual clients. A toxic "bot"ulism paralyzes Netcentric firms. But that's the worst case. Smarter organizations will recognize that, in practice, network media increasingly make deception and productivity more like partners than like rivals. That's hardly a happy conclusion. But it's one that recognizes what really makes people and their organizations run. Honest. Would I lie to you?

MICHAEL SCHRAGE is a Merrill Lynch Forum Innovation Fellow and a research associate with the MIT Media Lab. He may be reached at michael_schrage@fortunemail.com.