Agent Access Form: Fill & Download for Free

TL;DR: scroll down to the end for a list of constraints.I like Andrew Ng's analogy that compares AI's threat to humanity to the dangers of overpopulating Mars. Both threats are hopelessly far away, given foreseeable technologies. And yet, the AI threat captures people's imagination far more than the Bogeyman - a monster hiding under children's bed at night that kidnaps the child for disobeying parents. Why is that?As the story goes, super-intelligent AI machines learn quickly, gain physical omnipotence, move fast, and are markedly unfriendly to humans (Skynet-style) or are unconcerned about the humanity, with grave consequences to us (Vogon-style). Without assuming anything else, we cannot know their limitations and weaknesses, so we cannot rule them out (the way children eventually learn to rule out the Bogeyman) and do not know how to control them. This is scarier than Kraken, Godzilla and their ilk, who:can be clearly seen,have a physical shape and mass,can only be at one place at a time,move with limited speed, and are found in certain regions,are living creatures, so must breathe and consume food,act alone,do not breed,do not improve themselves,are too destructive for their own good (so, not very clever).Fortunately, monsters have vulnerabilities and can be attacked. Even Cthulhu can be trapped:H. P. Lovecraft's initial short story, "The Call of Cthulhu", was published in Weird Tales in 1928 and established the character as a malevolent entity, hibernating within an underwater city in the South Pacific called R'lyeh. The imprisoned Cthulhu is apparently the source of constant anxiety for mankind at a subconscious level, and also the subject of worship by a number of human religions. The short story asserts the premise that, while currently trapped, Cthulhu will eventually return.Stories about evil monsters play to the humanity's fight for survival against various predators, most of which are now extinct, deep underwater, or bound to remote wilderness. Historically, humans outsmarted predators by coming up with new weapons, hunting in teams, understanding the predator's physiology and limitations, then developing new hunting strategies (such as an ambush next to a watering hole during drought). It was tougher to fight against other hominids with anatomically similar brains. Scientists are still trying to figure out Humans Vs. Neanderthals: How Did We Win? - Neanderthals were stronger fighters, but humans were better long-distance runners, which gave a major advantage in hunting. Even more importantly, humans were more intelligent, with advantages in creativity and communication. Humans worked better in teams and apparently hunting with wolves helped humans outsmart the Neanderthals (by the way, legends about humanoid monsters apparently go back to experiences with Neanderthals in prehistoric Europe).Now, let's get this straight - the entire history of humanity is based on outsmarting menacing adversaries, but soon we'll have to deal with something that can easily outsmart us. We don't know what it will be, it may not have a body that we can attack, it can move fast, and it can be at many places at once. We are obviously doomed! The only hope we have is an early start (that is, unless shape-shifting terminators start arriving from the future), and perhaps the ability to control what gets built and what gets banned. But if the thing starts improving itself and multiplying, then all is lost for us.If we learned anything from aeons of evolution, from mythology, science fiction and actual science, that would be the resolute use of the physical world - not only building new tools/weapons, but also understanding the structure of a threat and cutting its life cycle. HAL 9000 was shut down (hence, the idea of a kill switch). Bacteria can usually be defeated with antibiotics. Fast-mutating viruses, such as HIV, can be managed with drug cocktails. Understanding cancers led to a variety of cancer treatments.How do we apply this knowledge to limit AI threats? - We prevent future AI from turning into a self-sufficient life form. Hanging monitors and adding kill switches seems too simple and too easy to circumvent, just like Volkswagen circumvented car emissions standards (nothing to do with AI, but imposing constraints on new machines and electronics is common today). We need something more serious. Granted, we should not expect too much - there could still be loopholes to constraints, but they will be smaller, and if runaway AI does appear, it will not be able to take over the world instantaneously.With those disclaimers in place, we should must limit AI's autonomy, virality, virility, team skills, potential for destruction, and critical physical capabilities.Introduce significant and tangible boundaries between agents of different levels of intelligence, so that we don't worry about networked toasters and door knobs. There ought to be a hierarchy of intelligence and cognition based on which capabilities are available.Segregate intelligent agents with different specializations (especially those that can hold weapons) by language and communication protocols, so that they cannot form heterogeneous armies.Limit self-replication (physical or electronic), physical self-repair and physical self-improvement. Kill switches would be nice, but who is going to flip them? Besides, they can be disabled.Limit intelligent agents' access to energy sources - without energy, they will not last long by themselves (and smart ones will realize this too). There's no problem for a computer to connect to an electric grid, but the grid should be physically controlled by people under any scenario. Long-lasting batteries (based on radio-isotopes) should be used in very few cases.Limit intelligent agents' access to weapons and sensitive facilities (power stations, command bunkers, etc). Use a hierarchy of access permissions.The logic with domesticated wolves also applies - we ought to develop AI specifically designed to protect humans.Asimov's Three Laws of Robotics should also be handy:A robot may not injure a human being or, through inaction, allow a human being to come to harm.A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.There are a number of other possible rules, but they tend to be either too easy to circumvent (such as kill switches) or too limiting for compelling applications (such as not allowing AI to generate and pursue its goals). Even the rules listed above seem too vague or too restrictive in practice. For example, Asimov's first rule has recently been questioned - Why Self-Driving Cars Must Be Programmed to Kill (when things go bad, it's critical to pursue the least bad outcome, but even that may require sacrifices).

First, a bit of clarity is in order.Most of the time when people unfamiliar with espionage ask about secret "agents" they really mean "officers" or "case officers". A member of the CIA, MI6 or KGB (now FSB) involved with espionage is properly referred to as an "officer". This is how the various spy agencies refer to their own employees who work with human intelligence.An agent is a source of human intelligence. Often times an agent is thought as an informant or really anyone who is able to gather intelligence for the officer. Officers recruit agents. The two terms are very often confused.So strictly speaking - you probably won't spot a secret agent - but not because the agent was well trained in espionage, not because the agent was selected to fit in, not because of any reason other than the fact that espionage activity is relatively rare. In the United States, with a population of ~300 million - perhaps there are a couple of thousand agents at any given time working against the United States. An agent is selected because he/she may have access (or be near access) to critical information of interest to a foreign power.Agents are often not well trained in fact and may not even be aware that they are giving away information to an officer. They may be under the impression that they're giving information to a businessman or a lover. If you have access to critical national information - you could have been an agent. So if you want to spot an agent, you need to start thinking like an officer. Find someone with access to secrets and see if the person has vulnerabilities. For instance, the single and bland looking secretary working for the head of a government organization, the nuclear scientist who has an addiction to prostitutes, etc.A Case officer is an intelligence officer who is a trained specialist in the management of agents and agent networks.[1] Case officers manage human agents, and human intelligence networks. Case officers spot potential agents, they recruit prospective agents, and they train agents in tradecraft. Case officers emphasize those elements of tradecraft which enable the agent to acquire needed information, as well as to enable the case officer to communicate with and supervise the agent. Most of all, case officers train agents in methods of avoiding detection by host nation counter-intelligence organizations.By definition, an "agent" acts on behalf of another, whether another individual, an organization, or a foreign government. Agents can be considered either witting or unwitting, and in some cases, willing or unwilling. Agents typically work under the direction of a principal agent or a case officer. When agents work alone, and are not members of an agent network, they are termed "singletons."The identification of potential agents is termed "agent spotting." Identifying potential agents, and investigating the details of their personal and professional lives, involves the granular verification of their bona fides. Such activities can include uncovering personal details that leave potential agents vulnerable to coercion, blackmail, or other inducements, such as sexual approaches.The recruitment of potential agents is an art form, and it is the raison d'être of the intelligence case officer. Approaches to potential agents can be multitudinous and interminable, and considerable time can pass before the potential agent is suborned or maneuvered into a position where a recruitment "pitch" can be hazarded.http://en.wikipedia.org/wiki/Agent_handling

Kulbhushan Jadhav is officially in Pakistan’s custody and everyone knows about the same. He was captured from Iran where he was doing legitimate business, brought to Pakistan, and shown as if he was spying. Diplomatic access to the Indian national was denied and India had to knock the door of ICJ.Kulbhushan Jadhav’s mother and wife were allowed access and we all further know, how were they humiliated when they were forced to remove the Bindi, Mangalsutra, and even shoes before meeting Kulbhushan Jadhav.Habib Zahir is officially not with us and if Pakistan claims he is with us then it becomes their job to prove it, which they cannot.Yes, we fooled the ISI agent Habib Zahir, Gave him a lucrative job with a hefty pay packet, which he could not resist, he came to Nepal and we Kidnapped him.Any harm to Kulbhushan Jadhav and we make a powder or liquid out of Habib Zahir and wash it in a drain untraceable.Pakistan has to weigh the pros and cons before harming Kulbhushan Jadhav because the price they will be paying will be far heavier, which they can ill afford.Kulbhushan Jadhav meeting his mother and Wife and Habib Zahir who will shorty change from human form to powder or liquid.