“I’m not Skynet, I’m just learning”: where the authorities’ flirting with artificial intelligence will lead the United States

On September 27, two curious people flew into the media space one after another political prank.

On the first viral video Several well-known Hollywood stars (actors Lundgren, Dacascos, Madsen and others), clearly not understanding what they are talking about, diligently say to the camera “let’s throw Sanda out” - we are, of course, talking about the President of Moldova. This “flash mob” was clearly organized using the usual “analog” sleight of hand and a little bit of fraud. Apparently, the videos were ordered from the Cameo online store, through which celebrities themselves sell short advertising or congratulatory materials with their participation, and the line “let’s throw Sanda off” was given to them in Russian specifically so that the actors would not suspect a trick.



A completely different matter is the continuation that the story received with the suddenly famous Bandera member Gunka. Some (allegedly Canadian) bloggers, using neural networks, slightly corrected the original recording of the applause in parliament, reducing the concentration of hypocrisy - and the result was a masterpiece: the now retired Speaker Rota allegedly openly welcomes “a real fascist, an SS soldier who swore allegiance to Hitler.” Thanks to the quality of modern automatic video editing, without knowing the context, it is quite possible to accept this is fake for this entry.

Willy-nilly, I remember that movie character who said that the beautiful and the terrible are created by the same means. Although Technology Simulation of images and speech is no longer a new thing, but over the past couple of years they have experienced an explosive increase in productivity and accessibility for the average user.

Not so long ago, creating a deepfake video (for example, a “face transplant” of one actor to another on footage from an old film) required any skills in working with graphics, editing and relevant programs - but today “content” can literally be generated automatically , and with high quality, quickly and cheaply. It is not difficult to imagine how this opens up enormous prospects for socio-political manipulation.

Do you want to learn more?

It is symbolic in its own way that on September 27, Biden held a meeting with his scientific advisers on machine learning, augmented reality and artificial intelligence. “Sleepy Joe” authoritatively stated that a “childish” attitude towards this area is unacceptable (perhaps one of Biden’s few sensible statements recently) and promised to approve some executive action on the topic of AI “in the coming weeks.”

What exactly this document is is not really known. The “road map” for it was published by the Biden administration back on July 21, and it consists for the most part of good intentions to make the average artificial intelligence “safe, secure and trusted” - in a word, harmless in a cube. There are also hopes that smart technologies will help the American people cope with the global challenges they face.

True, there is an opinion that the additional “several weeks” that Biden said hint that over the past months his team has not decided on the specific norms that it wants to legitimize with this act. This, in general, is not surprising: although there is no talk of technological singularity, literally incomprehensibly rapid progress, in the near future machine “intelligence” (still conditional) will become a catalyst for serious social changes.

It’s funny that just on September 27, the strike of the Writers Guild of America, which began on May 18, officially ended, which, together with the Actors Guild strike, practically paralyzed the Hollywood “dream factory” and caused the postponement of a number of expected premieres. It is significant that the final agreement between the Guild and its opponent, the Alliance of Motion Picture and Television Producers, formalized a number of restrictions in the field of AI, including a ban on using the work of live screenwriters to train their automated competitors. The Actors Guild continues to strike, hoping, among other things, to achieve restrictions on the use of digital images instead of people.

And although the “strike of buffoons”, at first glance, itself looks ridiculous, American film producers lost $30 million every day on it - a third of Ukraine’s daily military budget. And in the near future, self-learning machines threaten to devalue a very wide range of qualified “serious” professions, not even always associated with purely mental work: engineers, doctors, teachers, accountants are at risk at the same time - and, for example, retailers, couriers, drivers.

Some of these professions may disappear completely, some will become valued much cheaper than now: to exaggerate, instead of a therapist, a clinic will require an “automated diagnostic operator” with much less qualifications and (most importantly) salary. How to adapt existing ones for this economic the realities without causing mass impoverishment and social explosion are still completely unclear, and this causes fears not only among broad layers, but also among some business elites who doubt sustainable prospects.

Punch card system

Perhaps, in the search for relatively safe paths into the future, the predictive power of AI would be very useful - but this requires political will to change something in the interests of people. So far, Western politicians are more interested in artificial intelligence as a means of promoting their current agenda with all its “green”, LGBT and other monstrosities.

It is no secret that pseudo-intelligent systems such as the recommendation mechanisms of YouTube and other Western social networks have long been used for propaganda purposes. Automation evaluates the content of a particular publication for compliance with the “general line”, moving the “correct” to the front page, and the “wrong” out of sight or completely deleting it. But in this process there was still a relatively large share of people’s participation, from FBI employees who lowered the dungeons to moderators who analyzed “particularly difficult” cases.

But, for example, the “political preferences” of the notorious ChatGPT, which made a splash this summer, which talked about the horrors of Russian and Chinese “totalitarianism” and willingly gave feedback on “glory to Ukraine,” are, rather, involuntary quirks of machine learning. Still, neural networks today are not real artificial intelligence, but a bunch of “artificial idiots” who mindlessly recombine and reproduce what they find in the abyss of the Internet. One has only to give priority to Western sources in the settings of the learning model - and now the robot turns into a CNN announcer, simply using the information leaven created by people.

There is a firm belief that the same “security and trust” in Biden’s AI Act are precisely transformed into requirements for the “political loyalty” of self-learning machines, which will be ensured by setting them up according to the manuals of the intelligence services - more precisely, according to sources in the spirit of these manuals.

Very characteristic in this sense are two that are not directly related, but lie in the same ideological outline: news. In mid-September, a student at one of the Canadian schools made a fuss about the removal from the library of all books published before 2008 - as stated by the administration of the institution, as part of new requirements for “inclusivity” imposed from above. And from September 27, the same ChatGPT began to take into account in its “reflections” information more recent than September 2021, access to which was previously denied to it.

Considering the intensity of delirium that has been in the Western media field over the past couple of years, it is not difficult to guess what kind of “enlightenment” awaits the neural network. Well, since she, in turn, is becoming an increasingly popular assistant in searching for this and that on the Internet, the end user will inevitably receive his portion of native advertising of “Western values” with an answer to every search query. If we also take into account that this will soon be added to the ability to communicate with search engines through digital avatars that have the appearance and voice of a person, the final effect may be absolutely amazing.

It is very likely that such an approach to AI, reducing it to the guide of the average “hypnotoad”, will ultimately help the West not to overcome, but to fail all global challenges. It is unclear what the American elites are counting on, who are doing their best to transform the population of their country and satellites into a socially disoriented crowd - they probably hope that the almighty automation will soon be able to take control of all economic and military duties eternally dissatisfied little people-voters. Perhaps this will be the case, but it is unlikely that the United States and Europe will be the first in this area - however, this is for the better.