It started with robots.
In the first part of this 3-article series about artificial intelligence in science fiction vs. real-life AI, we talked about robots — both real and as depicted in sci-fi starting in the early 20th century. Then we visited advanced AI like the mutinous shipboard computer HAL 9000.
But that’s just fiction, right? When does it start getting real? We’ll conclude with a brief look at today’s robot police and computer-assisted crime prediction.
It turns out RoboCop is real … sort of. Leave it to Dubai to deploy the world’s first robot police officer. This life-sized robotic patrolman, put into service earlier this year, can collect evidence and identify wanted criminals as it patrols the busiest parts of the city. The police robot streams video back to the human police and can also be used to contact a live officer. Dubai hopes for its police force to be 25% robotic by 2030.
Minority Report and Predictive Policing
Where AI really makes the move from fiction to fact is in today’s predictive policing.
Not only is The Matrix a modern-classic sci-fi film, it’s also a London gang database. The Gangs Matrix lists several thousand suspected gang members living in the city — using data based on violence, criminal offending and gang membership.
Philip K. Dick’s “precrime” division in the sci-fi short Minority Report – later developed into a feature film starring Tom Cruise — uses the organic brains of three plugged-in “precog mutants” to determine where, when and by whom a crime will be committed, and preemptively lock up potential offenders. This isn’t strictly artificial intelligence, since the precrime knowledge is organically-based — but that knowledge would be worthless without the technology that channels and processes it.
“In our society, we have no major crimes, but we do
have a detention camp full of would-be criminals.”
— Minority Report, 1956 by Philip K. Dick
Nobody’s really using mutant humans to predict crime. They’re using data, and loads of it, including databases like the Gangs Matrix. Today’s big data is collected, refined and analyzed to find patterns in criminal activity.
PredPol software uses three data points — crime type, location and time — to provide law enforcement agencies with customized predictions for when and where crimes are most likely to occur. After PredPol was introduced, cities including Los Angeles and Atlanta reported drops in crimes like burglary and vehicle theft.
To show predictive policing techniques in action, the makers of the documentary PRE-CRIME traveled to Paris, Berlin, and other cities around the world with an eye toward possible abuses. Predictive policing and databases of information about potential criminals raise the ethical issues of profiling and potential bias, but they make strides toward the goal of preventing crime before it happens.
Where Fact and Fiction Meet … and Don’t Meet
Artificial intelligence in science fiction is beginning to merge with reality. The best sci-fi is based on real science, and it’s becoming more plausible all the time. We don’t have flying cars like The Jetsons, but self-driving vehicles are part of today’s reality and Roomba is our robot maid. Our vehicles don’t chat with us like Knight Rider’s KITT, but voice commands control activities like activating GPS, selecting a radio station and making a call.
AI is used in medical imaging, weather prediction, oil and gas, manufacturing, and many other industries. Not only has it enabled new discoveries, it’s becoming a mandatory tool for both commercial enterprises and scientific researchers.
One concern about artificial intelligence is that “smart robots” will take too many jobs away from humans. Many repetitive processes have already been automated. What if computers, which can already outperform us on many tasks, develop the ability to take on more complex jobs that require judgement and abstract thinking?
Whatever happens in the future, AI has undoubtedly made human lives easier right now, allowing us to be more productive in ways machines can’t match.
The big question is what could happen if — or when — machine intelligence surpasses human intelligence. This is known as the singularity: the point at which self-improving AI learning cycles could leave humans behind and create a machine superintelligence. Tech entrepreneur Elon Musk says AI could be “our biggest existential threat.” And Stephen Hawking told the BBC that “the development of full artificial intelligence could spell the end of the human race.”
But today we can relax. Artificial intelligence isn’t ready to take over. Machines are skilled at crunching big data and making inferences based on extensive training, but they aren’t (yet) capable of complex thought and analysis. Today’s AI is an important tool that supplements our skills and enables us to discover hidden knowledge. Because of AI, we’re making unprecedented advances in science and industry.
What do the robots think of all this? So far, only what we tell them to think.