Google, Baidu and and others are working hard at visual recognition software. Baidu just reported it now has an AI system that recognizes images better than humans. It can distinguish 1000 categories, so that would be trees, houses, people, dogs etc. This ability is implemented such that it could be used as a service, meaning if you fire an image to Baidu, it will tell you what category it is.
Already this has consequences for our daily lives. It is highly likely that Baidu will now be able to categorize websites by their image content, and for instance block any sites with to much nudity or other offensive material. But these recognition systems can be trained in specific fields, they can process 2d or 3d info, this doesn’t make too much of a difference. This means they can not only recognize but also validate a real situation, if this is offered to them in image form.
Now imagine you want to run a hospital, and you have cleaners for the rooms, and you want every room to look tidy after they are done. You can either have someone check the rooms, or you can ask the cleaner to scan it after they are done. The scan gets uploaded to a Baidu image analysis system, and you get your answer about the quality of the work.
This simple principle works in many different places. A lot of what we do is judged by visual cues. And Baidu type systems can be trained in a specific area of expertise, so for instance to look at seams in clothing (there’s even a system that recognizes criminals in camera footage by the seams of their jeans, but that is a segway). Camera observation systems can now become overseers. One can not only look for possible burglars, but also whether the place looks like it should at any given time.
How would it be if we tie this functionality to access to our cash, or the ability to get fuel at the gas station. In the coveted cashless society anyone can now be asked to document their actions, have them ‘approved’ by Baidu type recognition systems, or find themselves be cut off or punished (lets not use euphenisms here).
Work benches can have cameras just like factory conveyor belts so that someone at Foxcon can lay his Iphone under the cam at every step of the assembly process for a green light and 5 sleep credits. AI only needs good recognition systems, because we already have a super control system for humans, a bank owned money supply. Money is the gateway to everything we need, or banks are trying to make that so, and with that we are imprisoned even before these systems become implemented.
The surveillance state is one that exists because of fear and scarcity, the scarity creates an elite which then uses fear to justify surveillance. The primary scarce thing in this mechanism is fossil fuels. A surveillance state would not come about if everywhere you go around the world people would meet their local needs with local renewable energy. There would be no need to go elsewhere and take or ‘share’ oil as the west does in the Middle East, People would not become terrorist except in a rare case of insanity. People would not be afraid because they would know what their live depends on, and it would not be money because money would be virtually elimintated (as most things would be free and produced automatically see: the Roboeconomy).
Fossil fuel (natural resouce) scarcity drives surveillance and the emergence of a slave class
The Baidu system may in future include recognition of smell. It can already easily include recognition of audio, something easily adapted for making people chant affirmations in the morning. But these captivating opportunities to totally dominate a work force with no human effort are misguided in the sense that our competition for cash is one born out of an unnecessary scaricty of energy, one we can eliminate with renewables.
Let’s hope we achieve a state where renewables are overabundant, and we don’t need to have human slaves work for food and water credits in FoxConn factories until they ask to be euthenized. The fossil fuel /banking elite will certainly like to see it, and your weapon is to drive the building and realization of more renewable energy sources, and replace the remote factory worker with a local who you can check is happy.
Content Surveillance
It is easy to imagine that one develops a ‘content observer’ (content as in happy) application that can be used with any image feed in a fixed place. When things are as they should be, the system is content, when it is not it signals this and draws attention to it (of either humans or some other AI system). So town square with relaxed tourists : ok, slight altercation with between to men : Not ok, send sentries. Robots arrive at the scene and ask the fighting men to stand apart to be searched. Not happening? Taze them etc. etc.
These ‘content obeserver’ systems really do almost as much as a humans do, and they can of course be linked so that we get content areas, zones. This may however in our current economy mean that the people in the zones are doing what they should do, even if that makes them unhappy. This factor keeps depending on whether it is an elite that has to fight over resources or whether they are abundand as they can be.