Another CES, another hype cycle. This year, it's network-connected devices, which got their own interest group in the form of the Internet of Things Consortium.

Be afraid. The CES spotlight tends to reveal the absence of innovation rather than the next big thing. Last year, we had the year of the ultrabook. But it wasn't. PC sales declined 7% in 2012, according to IDC. Before that, it was the year of the tablet, but only Apple iPads were really selling. CES 2010 brought us the year of the 3-D TV. I thought it was a bad idea at the time and now it seems others share that view.

But the Internet of Things -- making objects Internet-addressable and responsive -- is a better bet than 3-D TV. It's inevitable and is already manifesting in many ways. It's evident in traffic data aggregated from mobile phones and in the way TiVo customers can program their DVRs from the Web. It's evident in the way Nest customers can read and alter their thermostats online. It's the next big thing because it's already here.

But in the rush to connect physical objects, we risk ignoring the return on investment and the cost to ourselves.

[ Tech fans see a lot to like at CES. And then there are the things we love to hate. Check out CES 2013: 5 Dumbest Ideas. ]

Let's start with an easy target: Samsung's Smart Refrigerator. It's not smart, and at $3,700 it's not very affordable. It's a fridge with an attached touchscreen, something that could be accomplished with an iPad Mini and tablet mounting hardware ... if that were preferable to just placing your favorite tablet computer on a nearby counter.

Samsung's suggested uses reveal just how useless this product is: "Check the morning weather, browse the Web for recipes, explore your social networks or leave notes for your family -- all from the refrigerator door." None of these tasks enhance this refrigerator's primary function, storing perishable food. All of these tasks would be better accomplished with a smartphone or a Post-It note, or through the miracle of human speech.

It's possible to imagine how a smart refrigerator might be useful, if it could, say, tell you to buy more milk as you approach a grocery store on your way home. But it would take a fairly sophisticated system to assess the remaining volume of milk in the fridge and to communicate that information at the optimal time.

Better to rely on a smart person than a smart appliance. People are extraordinarily well-equipped for assessing what's in a refrigerator -- hey, we're low on milk! -- and re-stocking it as required, without a network connection. Consumers do not need the inventory tracking capabilities of Walmart.

Moreover, people actually get something out of running their own lives rather than being mere endpoints for network notifications. By outsourcing attention to sensors connected by the Internet of Things, people risk surrendering their core competency: engaging with the real world.

What's more, we should be wary about fetishizing data. If you need data to make decisions, by all means gather your data and act on it. But detailed graphs of energy usage, courtesy of your networked Nest thermostat, won't change whether you want more heat because you're house is too cold. A $50 sweater is probably a better energy savings investment than a $200 Wi-Fi-enabled thermostat.

This is not to say the Internet of Things Consortium is a bad idea. Quite the opposite. But as group chair Jason Johnson explained in a phone interview, the IoTC isn't in the business of creating networkable things, a business sure to produce more disasters than diamonds. Instead, it aims to promote data sharing and API development and to analyze market trends.

"When you have one company doing one thing really well, it's very important that others can interact with that product in a two-way manner, where said company could not do that on its own," said Johnson.

Johnson cites Sonos and Logitech as examples of companies making devices that play well with other products. This is an implicit rejection of companies that seek to be platform gatekeepers. Networking things should rely on open standards rather than license-based connectivity. And to the extent the IoTC advances that agenda, it should be appreciated.

But just because you can network your refrigerator, clothing or fork doesn't mean you should, not only because human oversight is often much more effective than mechanical monitoring, but because some of the potential benefits of Internet-connected objects obscure even greater costs.

Have you ever jaywalked, exceeded the speed limit, lingered in a parking space beyond the time you paid for or broken rules in a minor way? Would you want to pay for these infractions? When everything is connected, nothing can be hidden, particularly when exposure translates into revenue for government agencies, businesses or entrepreneurs. Would an insurance company pay to know if its health insurance clients keep only unhealthy food in their refrigerators, so it could mitigate a higher risk of health problems with higher premiums? Count on it. This is the world the Internet of Things makes possible.

It doesn't have to be that way. Having a car with Internet-accessible systems shouldn't mean that law enforcement will have a backdoor to track you or to shut your car's engine down remotely. It shouldn't mean that your driving data will forced to testify against you. It shouldn't mean that air quality sensors in your home or chemical sensors in your sewage pipes should notify the authorities when controlled substances are detected.

Be afraid of the Internet of Things, because things are not the same as friends. But be open to things that can be tamed.