TECH | Oct 12, 2017

Name of the medicine: Artificial Intelligence

If Artificial Intelligence were a medicine: read the warnings and instructions for use


Artificial Intelligence or AI is, according to Wikibooks, “the ability of a computer to perform functions and reasoning typical of the human mind”. More specifically, according to the most quoted definition on the internet, it is the set of “theories and techniques for the development of algorithms that allow machines to show intelligent skills and/or activity, at least in specific domains. Intelligent activity is taken to means the ability to extract guidelines from previous knowledge to be used to solve new problems, problems that the system has never faced though it may have dealt with similar in the past”.

The original active substance of the drug General Artificial Intelligence (GAI), has long been at the center of the research activities of the dedicated pharmacists, as well as of numerous science fiction authors: it consists in the ability to fully replicate human intelligence. Given the complexity of synthesizing this active substance, it has been decided to begin prescribing equivalents of minor complexity. Specifically, the most popular generic drug is Weak Artificial Intelligence (WAI), namely that artificial intelligence dedicated to analyzing and/or solving specific problems, sometimes computationally too complex to be analyzed by the human mind. This is the case, for example, of the “Deep Blue” drug, used to beat chess champion Kasparov. Or of many other widespread applications that are often not known to be based on this active substance (see the paragraph on “Why it is used” in this information leaflet).

In order to complete the range of equivalents of WAI, we cannot fail to mention Strong Artificial Intelligence (SAI) which – like GAI however – has not yet been synthesized due to the complexity of the molecule. SAI is that Artificial Intelligence which, while not fully replicating the human mind, is supposed to be able to reach self-consciousness.

The medicine, even though authoritative sources do not report it, is capable of enriching and extending human capabilities, skills and potential to the extent that it can be used – in selected but rare cases – as a remedy for natural stupidity.


In defining the WAI equivalent in particular, Artificial Intelligence is used in cases where it is possible to automate and optimize actions or services typically characterized by repetitive actions which, starting from input data and correct responses, lead a machine to do the work of a human.

It has already been extensively tested successfully in cases of virtual assistance aimed at supporting the user who asks for something and can have an answer; in the suggestion of “things you like” related to user behavior, for example on social networks or search engines; in the fight against crime with analysis of images captured by video circuits in cities or places of public interest; in the analysis of data intended to foresee customer purchases; in home automation or smart homes; in the prevention of fraud, which, by analyzing the habitual behavior of the client, is able to intercept possible potentially harmful exceptions; in video games; in the much-feared and debated autonomous driving of vehicles and even in journalism with automatic systems for writing short news reports. In this regard, a recent study has found that – for certain categories of articles (sports, finance, weather forecasts) – a representative sample of U.S. readers preferred the articles written by a computer to those produced by “human” journalists.

Distribution of the medicine is under way on an experimental basis (perhaps better said, has been announced, complete with a working table) in Public Administration where, according to AGID, the challenges to be overcome are ethical, legal, technological, cultural and security in nature but are also linked to the need for specific skills, the availability of interoperable and usable data, and social impact.


Do not use the medicine if you want to keep a situation, process, company, or body unchanged.


The availability of Big Data, the mental openness of people involved in drug testing, the willingness to welcome innovation and the willingness to break consolidated patterns can deeply change the effects of AI, and go on to increase its effectiveness.


How much

Artificial Intelligence must be taken in small amounts, especially in the initial phase of experimentation. The doses and scope of application can be increased gradually once the positive effects of the drug have been evaluated. Contrary to popular belief, AI is not harmful to health (except in the cases described under “When not to be used”).

When and for how long

AI should be taken regularly, with a low initial dose that can change over time. The benefits of the drug are perceived in the medium to long term, so you cannot expect instant improvements or even within a few days. The efficacy of the active substance intrinsically requires a long time, which is necessary for supporting the continuous learning process of the system for improving its effectiveness. If taken correctly, it creates dependency.


Artificial intelligence can be taken at any time of day and night, week, month or year.


If AI becomes harmful, it is advisable to suspend immediately.


No interactions with other drugs are known. The effect of Artificial Intelligence is increased by the presence in the body that makes use of Big Data bases which act as catalyst for action of the active substance. At present, attempts are being made to understand the interactions of the drug with elements such as the concepts of work, ethics and privacy.


The main undesirable effect, which has been theorized but not yet verified in the case of adoption of GAI, is the physiological phenomenon of “technological singularity”, in which machine intelligence comes to surpass that of humans. Researchers are divided on the concrete effects of this singularity: some believe that the machines will free humans from work and others think it will reduce them to slavery. Artificial Intelligence can lead to nervousness and panic attacks in those who fear the optimization of human resources and streamlining of processes. But also in those who have to argue with not so smart Artificial Intelligences: this is the case for many automatic responders and – at present – almost all virtual servers installed on smartphones. In such cases, discussions with them may lead to particularly grisly results for smartphones (reported cases of smartphones thrown out of windows). Frequent cases of anxiety and phobia have emerged in those who rely on titles that evoke fantasy science scenarios of robots taking control of the world.

AI systems may activate brain reward mechanisms (such as in video games or on social networks) in the event of prolonged intake, which could give the impression that it is impossible to give up the drug; in other words, dependency.

Sonia Montegiove – Stefano Epifani