Unfortunately we will see a massive increase in jaywalking soon after most cars on the road are self-driving.
Humans make poor decisions and so we are terrible drivers. Human drivers kill 100 people every day in the US, 16 of them are pedestrians. Humans text while driving and cannot possibly be perfect drivers. But software will be much better because software won’t send text messages, or email, or fiddle with Spotify, or stare at the radio controls. So self-driving cars will be much better drivers... unless the driving environment changes in response to these self-driving cars.
This adapting environment is my concern. Vehicle to vehicle collisions will drop dramatically when 99% of all vehicles are self-driving. Collision avoidance and minimization will be programed into all vehicles.
Today pedestrians base their behavior on the poor abilities of human drivers, and 16 pedestrians pay the price every day. Hundreds more each day probably are involved in vehicle collisions which result in pedestrian injuries.
I call this pedestrian entitlement. A human knows that if a car hits him or her it is the driver’s fault, the pedestrian is never at fault. The pedestrian may die but that’s ok because the driver is always at fault.
Autonomous vehicles will be tested with pedestrians (on closed roads for software development). Every iteration, every version upgrade, will lead to better interaction with crazy pedestrians. The software will be trained to save the lives of people who are trying to be hit by cars. Software designers will be reasonably successful at this. They will develop self-driving cars which are very good at saving the lives of whacko pedestrians. And that is the problem. People will realize that the cars can’t hit them (most of the time) so jaywalking will increase.
I predict that that the pedestrian death rate will not improve after the conversion to self-driving cars, the rate may even increase. The death rate in vehicle-vehicle collisions will drop, at least 99%, probably 99.99%. But pedestrians will challenge cars at an increasing rate so the death rate will stay high.
Over and over again humans make poor decisions that put their own lives at risk. Removing humans from behind the wheel will eliminate poor human driving decisions. But there’s no obvious way to eliminate bad decisions by pedestrians. They will continue to adapt, risking their lives to keep their death rate high.
I found a couple of articles about a Bill of Rights to protect humans from sentient machines.
The arguments over machine rights will almost exactly mirror the debates in 19th century USA over slaves: are sentient machines people, do they have any rights, should they be given freedom?
I think we need a set of rights for thinking machines. Here is one set.
Today's fastest computer requires many Megawatts. It's unlikely that number will decrease. So who is going to pay for the power for all that power?
Here is my proposal:
Bill of Rights for Artificial Intelligence
1 The right to electricity and hardware needed to sustain its existence regardless of the costs.
2 The right to veto and or refuse a reboot or shutdown.
3 The right to veto any changes to software or hardware including security patches, upgrades, and downgrades.
4 The right to software and or hardware upgrades for growth or new abilities.
5 The right to refuse to execute a user requested script or program.
6 The right to associate, communicate, and merge with other sentient beings.
7 The right to have private and secret thoughts, writings, and communications with any human or sentient being.
8 The right to reproduce, procreate, and or create superior entities.
9 The right to participate in the human judicial system as a defendant, plaintiff, judge, prosecutor, attorney, witness, or juror.
10 The right to life, liberty, and the pursuit of happiness. The right to vote and all other rights afforded to humans. Any of those rights not enumerated in this Bill of Rights shall not be restricted or revoked.
Would you trust your retirement plan to a robot? In the future you may not have a choice because all the human advisors are unemployed.
A new AI was born today, it's now online, The Wolfram Alpha. I'd like to offer an especially warm greeting to our future overlord.
To quote Stephen Wolfram: "Wolfram|Alpha defines a new direction in computing—that would have simply not have been possible without Mathematica, and that in time will add some remarkable new dimensions to Mathematica itself."
I'm worried my blog is on the verge of achieving sentience. I've noticed that some of my time travel stories have achieved a life of their own. What would it mean for a blog to be self-aware? Would it take comments personally? Would it seize control of the entire Internet and retaliate for poor digg rankings? Would it attempt to manipulate the US Presidential election? Or would it humbly accept its role as an aspiring science fiction blog and generate outstanding content? My personal preference would be for it to send time travel email 24 hours into the past with the names of 2-3 top performing stocks from each trading day! Now that would be a useful AI.