In aviation we have the ARP4761 Safety Assessment Process which is pretty much every aviation authority requires. Any use of AI in an aircraft will fall under that. Other industries where peoples' lives are at stake have analogous requirements. Thus if AI is part of a system the failure modes will be identified and mitigated.

As far as crimes are concerned, the fraud law is not concerned about what tools a perp. may have used to commit the fraud. It is still fraud.

So the question becomes "What AI specific harms exist?" Some science fiction writers have impressive imaginations and I suspect that their stories might be behind at least some of fear. It is obvious to me that much of the grief was easily avoidable. For example, given what we do to prevent the unauthorized use of nuclear weapons, why in "The Colossus Project" would anybody in their right minds give unchecked control of the nuclear arsenal to a brand new, untested AI? Having worked in aerospace for years I could find lots of faults in the design of the "Discovery 1" in "2001 a Space Odyssey". Why give HAL unfettered control of the suspended animation pods when the most it needed was fault monitoring so that it could raise an alarm.

If there is any AI specific regulation it should be "Thou shalt apply the 'Least Privilege Principle' in controlling what an AI can do." But then we have applied that to "natural intelligence" for decades.

I think, at the end of the day, there will be dozens of minor debates about how existing laws apply to AI. For example, the Fair Use of Copyrighted Materials has come up. People are allowed to read books and journal articles and apply what we learn to our work, but some authors and publishers think that it is not Fair Use when AI does something similar.

I suggest we all take a cue from The Guide: "Don't Panic".

Has everyone forgotten Sony Music vs Cox Cable? Sony won a billion dollars in that one.

There is more to it than federal research subsidies and politics. There is genuine fear. Fear does not need a basis in fact to be real.

A large part of the public is afraid of anything radioactive. It's invisible, has no smell, can not be felt or heard and it can kill you. It's the perfect bogey-man. I have found that facts don't usually matter when someone is truly afraid.

Then there's the Nuclear Regulatory Commission. The people there are afraid for their jobs. The result is paranoia in regulation. If they decide that high-tech pipe B might possibly be even slightly better than high-tech pipe A, they will mandate an upgrade for all plants still under construction, even if it means jack-hammering 12 ft of reinforced concrete to replace the pipes.

When you factor in the lawsuits from self-appointed watchdog groups it becomes impossible to estimate either the total cost or when the plant will come online. With those in doubt it is impossible to determine whether the plant will be a good investment, and thus no private firm will take the risk.

There is a very real possibility that ISP's will be required to enforce copyright laws in the same way that convenience stores are required to enforce age limits for alcohol and tobacco. ISP's might also lose the "safe harbor" provisions and become "accessories" to the actions of their users.

If either of these possibilities becomes law the ISP's will be required to shut down IP infringing traffic. So it could be evidence that ISP's are looking for a way to comply with such laws should they be passed.

It would not be the first time that the U.S. Congress has put a deadline on a technology which did not exist yet.

"No man's life, liberty or property is safe when congress is in session."

Another feature is software simplicity. This may not be an issue for the laptop, but it is an issue for the embedded system. Or it can be during development.

USB is a complex protocol which requires a fair amount of code and data structures to support. A serial port on the other hand requires less than a page of code (in it's most simple form). The result is that when a system crashes, a serial port has a much better chance of being operational than a USB interface. Many systems with serial ports are designed so that a break signal on the line will interrupt the processor from whatever it's doing and send it directly to the debugger. When you can examine the entrails it is much easier to divine the cause.

Of course it is possible to design a bit of hardware which looks like a USB serial port adapter to a laptop and a serial port to the embedded system. Even better would be a new USB interface which gives full access to system memory and processor state.

Presumably all DVD readers made for the next 1000 years will be backward compatible. Have you tried to read an 8-inch floppy disk lately? And they're only three decades old!

When the equipment for reading these starts to become museum pieces people will migrate the data to whatever the state of the art is at the time. Then these stone DVD's will last a long time in the landfill.

It does raise some fun things to speculate about though.

There are some ancient writings which no one knows how to read anymore. Will future archaeologists wonder what the microscopic pits in our coasters with holes in them are all about?

Will they suffer from data overload?

What will future archaeologists, with PhD's, think when they read what you, personally, wrote in a forum? Now that's scary.

This device sounds a lot like a non-contact polygraph. Now polygraph testing has been somewhat discredited of late, but that does not mean that it's useless.

It could become a useful supplement to the current search process by indicating who the agents should spend extra time on, or who needs to be checked by a more experienced screener.

Rail has been a popular environmentalist cure for traffic, pollution and fossil fuel use since at least the Arab oil embargo of 1972.

The issues which have prevented its universal adoption across the United States are still here.

1. Legal costs
2. Right-of-way acquisition costs
3. Construction costs
4. Traffic Disruption due to construction (an intangible but real cost)
5. Operating costs
6. Maintenance costs
7. Americans still want the freedom that cars give them.

Don't hold your breath on rail.

Detroit didn't come up with SUVs to dupe anybody. SUVs were popular because of their versatility, perceived sturdiness and their status.

The Big 3 have labor costs about three times higher than other auto makers in America. They also pay pension and health care benefits for about twice as many people as are currently working. In order to make a profit they had to sell large high-priced cars.

The high gas prices scared a lot of people away from SUVs for now, but what Americans want in their cars has not changed.

First of all, the automobile represents freedom. Freedom to go where you want, when you want. You are not tied to mass transit schedules and routes.

Americans still want cars that are status symbols. Even those who buy hybrids do so to show how much they care about the environment.

Americans want cars that are safe and useful. A family of five wants a car that can comfortably haul the family plus a couple of friends plus their stuff.

Since when do you have a reasonable expectation of privacy when you are visible from a public street?

I'm going to take a wild guess here: Some folks have never lived in a small town.