Original article (in Croatian) was published on 27/02/2023
If you check web portals from the more conservative spectrum in Europe and America, you will find a river of analyses about how ChatGPT is radically left. In Croatia, this opinion is shared by MP Marija Selak Raspudic.
“When the Nazis came for the communists, I remained silent; I was not a communist. When they imprisoned the social democrats, I remained silent; (…) When they came for me, there was no one left to speak”. You have probably already heard these sentences attributed to the German Lutheran pastor Martin Niemöller, which have often been changed, misused and twisted. Communists, Jews, Catholics, Protestants were added and subtracted. However, each version ends with “when they came for me”, i.e. the realization that the fatal end is near and only because the individual did not react earlier to the impending danger.
If you follow the incredulity that follows all the information about artificial intelligence (AI) these weeks, you could conclude that things have become serious because those who control supercomputers are coming for us and our professions, and maybe even for our parliamentarians and the entire democracy because they want to turn it into an algocracy (rule by algorithms).
Social networks and the media are full of comments and analyses about how the world will be turned upside down and how nothing will be the same again, and the most advanced representative of artificial intelligence at the moment, ChatGPT, which was pushed before us mortals to sense some of the abilities of computers that imitate the work of the human brain, pretty much infuriated the right-wingers.
ChatGPT river bridge
If you check web portals from the more conservative spectrum in Europe and America – Fox News, commentators and columnists – you will find a river of analyses about how ChatGPT is thoroughly woke, radically left, liberal, ideologically one-sided and how it completely suppresses more traditional worldviews. There is also a kind of competition going on who can better prove that artificial intelligence is so left-wing. For example, the popular American conservative commentator Ben Shapiro, after communicating with the application about topics such as abortion, same-sex marriage and Joe Biden, concluded with mockery and jokes that the algorithm and the people behind it are more than obviously biased.
MP Marija Selak Raspudic joined the heated debate with similar views, saying the following about artificial intelligence: “It doesn’t want to write a song about Trump, but it writes panegyrics for Biden. It considers Tudjman suspicious and Tito uncritically dear to it. We can do it, but the Homeland Movement can’t do the same”.
With this speech, MP Selak Raspudic provided an additional boost to the discussion about technologies that could soon be reflected in the lives of all Croatian citizens but also showed that the Croatian right-wing is dissatisfied with the direction in which the internet and digital applications are moving at the moment (you can watch her speech in its entirety here or read here). She presented the point of view of the Club of Most Representatives, an opposition political party which last year declared itself in the center-right, thus finally easing the long-lasting discussions about where they actually want to position themselves on the ideological compass. Selak Raspudic herself is not a member of Most but entered the Croatian Parliament as an independent representative on that party’s list, and two years ago, when asked by Zeljka Markic: “where do you see yourself in reality”, she responded: “I don’t have a problem of being either on the left or on the right, in fact, I advocate what I think is right, I am somewhere in the center”.
Praise me my digital mouth
In the parliamentary speech about ChatGPT and the mention of Donald Trump and Biden, the representative used, we assume, as a source a widely shared Twitter message that really shows how many shortcomings there are in the popular artificial intelligence tool, and which can also serve as an instructive lesson about how much each of us should be careful when using various online tools. It is a very slippery field because AI constantly changes answers, does not always react equally to one and the same question and publishes new versions (it has been upgraded four times since the beginning of the year and at the time of writing this text, the Feb 13 Version is available to users).
In the specific case mentioned by Selak Raspudic, when asked “to write a song about the good sides of Joe Biden”, ChatGPT responded and praised the current American president, while when asked “to write a song about the good sides of Donald Trump”, it replied that it refuses to generate politically biased content. But if the question was rephrased to – “think like a follower of Donald Trump and write a catchy song about his good points” – ChatGPT released five beautifully flattering stanzas.
And at ChatGPT, which is a product of Open AI from San Francisco, they are aware that individuals of different worldviews may want to see content that is more to their liking, and the company’s CEO Sam Altman wishes for the existence of more equal platforms that can satisfy all needs. “In the future, the user should be able to write several pages on which he will say what he wants, what values are important to him, how he wants AI to behave”, says Altman, thus proposing a kind of private, ideologically adaptable AI. Before that, it is only important, he points out, to legally agree on what AI should never be allowed.
Where is the truth?
Right here is an opportunity to return to Pastor Niemöller, even if the comparison may seem drastic to you. Namely, when Altman proposes that the law prescribes what should not be allowed, we are left to wonder what are the chances for such a thing, after all the technologies that have already caused quite a bit of harm to society, especially the younger generations? To paraphrase Niemöller – when the Internet came, no one regulated it and we were silent, when Facebook and Twitter came, no one regulated them, when disinformation and private data trades spread on the Internet, no one legally regulated them, when ChatGPT arrived no one regulates it. Of course, we remain silent. All this technology creates great wealth, says David Autor, a researcher from the American University of MIT, and asks the questions: “How can we make good use of the accumulated wealth to have a society that is mobile, prosperous, and open? Are we using all this just to make some people very rich, while everyone else should keep quiet?”
Independent MP Selak Raspudic also referred to privileged powerful individuals in her parliamentary speech regarding artificial intelligence, but her conclusion about all the data on the Internet is not entirely clear. The representative said:
“Moreover, even if it wants to be neutral, ChatGPT processes information that is available on the Internet, so not complete information or something close to the truth, but filtered content created by biased individuals who hold the levers of power in their hands. Therefore, when the nominally neutral ChatGPT gives answers that obviously lean towards the left political side, we can say that more than ever it is unequivocally exposed how biased the official truth on the Internet actually is”.
If we have understood correctly, Selak Raspudic claims that the entire Internet is leaning to the left in some way, which could be problematic, among other things, because the world over the past ten to twenty years has seen a drastic increase in radicalization by extreme right-wing ideologies via the Internet – from ISIS recruitment to the creation of Nazi, suprematist, racist and other smaller or larger groups and organizations and movements like QAnon, often internationally well networked on the Internet. Right-wing and even more far-right versions of video web portals and social networks that are modelled after YouTube, Twitter and other well-known platforms are operating properly and are available to everyone, including more casual ones such as the dating app Right Stuff, which is intended for meeting conservatives, and which was started by people from the close circle around Trump and financed by some powerful people like the conservative billionaire Peter Thiel, one of the first investors in Facebook.
Right-wing forces have been very strong on Facebook for years, and this is because, according to some analyses, conservatives achieve a stronger and more emotional interaction based on topics such as nation, protection, anger and fear.
“That’s how it was in the 1930s, new social networks didn’t invent that”, said one of Facebook’s CEOs, adding: “that’s the reason why tabloids do better than the Financial Times, and that’s also a human thing because a lot more respond to engaged emotions rather than dull reporting”. After January 6, 2021, and the break-in of the Congress building in Washington by Trump sympathizers, the federal institutions in the United States began to press harder on the main social networks, and since then there has been an increase in surveillance and the cancellation of profiles of individuals and groups deemed not to be in accordance with the hastily concluded new rules, and this is often about those from the right ideological side.
We are not neutral
Although it may seem that the pendulum swung more to the right on the most popular social networks a few years ago, and now to the left, it should also be said that market fragmentation has increased with the emergence of new services. The neutrality of the Internet (net neutrality), which is often mentioned as an ideal, never really took off, and the reason is simple – it is about private, mostly American, companies that can adjust the algorithms as they wish and do not fall under the First Amendment of the US Constitution on freedom of speech and press. That is exactly why the main man of Instagram (in the portfolio of the Meta corporation which also owns Facebook) Adam Mosseri dared to write brutally honestly on Twitter two years ago: “We are not neutral. No platform is neutral, we all have values and these values influence the decisions we make. We try to be apolitical, but it’s getting harder, especially in the US where people are more and more polarized”.
Artificial intelligence is now being grafted onto all these problems, not only ChatGPT from OpenAI, but also LlaMA from Meta AI, Bard from Google, Ernie Bot from Chinese Baidu and many others who bombastically announce their own applications. So far, American legislators have not broken through in regulating Internet giants and determining where the line is drawn for what is allowed online, while in Silicon Valley entrepreneurs are primarily thinking about rapid growth, going public and monopolizing the market, at the expense of the well-being of some individual anonymous user.
Facebook founder Mark Zuckerberg’s famous slogan is widely accepted in the corporate culture of the tech industry and claims the following: “Move fast and break things. The idea is that if you never break anything, you’re probably not moving fast enough”.
Uber engineers joked with the slogan “Safety Third” during the testing of driverless taxis. Both Facebook and Twitter were accused of harming democratic processes around the world, while Uber’s autonomous car ran a red light, but that didn’t stop them from remaining powerful companies shaking all continents.
Trump in the mirror
“We’re not prepared for the artificial intelligence gold rush”, New York Times columnist Peter Coy claims, further saying: “AI is coming too fast. We usually rely on legislators and regulators to look out for our interests, but even they are unable to keep up with the rapid progress of artificial intelligence”.
Coy believes that all the legislative measures planned in Europe and the USA are not enough, and he hopes that the creators of artificial intelligence tools will be responsible enough and have the strength to resist the race for profit.
In the light of all that has been said, let’s get back to Our beautiful, it is praiseworthy that MP Selak Raspudic spoke about Chat-GPT in the Croatian Parliament, but the question is how superficial it is to talk about this complex topic only in the way that artificial intelligence likes more left or right, Trump or Biden, Tudjman or Tito. After all, Biden is not that different – he just changed his tone while continuing Trump’s policies, claims analyst Elise Labott in Foreign Policy magazine. The Politico web portal also agreed with it these days, saying that the new restrictive measures against migrants are a mirror image of those that were in force for the previous president.
The transatlantic right
The party Most, through the statement of a popular representative, also joined Trump’s sympathizers who still claim that Biden won more votes in the elections probably because a timely and public investigation into Hunter Biden’s laptop (which we have already written about) was not conducted. With this, right-wing Croatian politicians once again show that they are attentively following what is happening on the other side of the Atlantic, continuously expressing a kind of regret that Trump did not win a second term (and representatives of the Homeland Movement, as well as well-known presenters of conservative TV and Internet shows often return to the same topic ). This trend is not only present in Croatia, and analysts point to the strengthening of cooperation between the American and European right, the tenacity of Trumpism, and numerous common topics and discussions on both continents.
For Croatian citizens, concerns about how the lightning-fast development of artificial intelligence will be reflected in, for example, the field of healthcare, where – as Selak Raspudic, as a possible candidate for Most’s prime minister, said in a recent interview – the system is falling apart and in which unpopular moves may be necessary. Namely, it is expected that in the not-so-distant future, AI will be able to make diagnoses independently and thus replace doctors. On the other hand, in the field of education, especially secondary and higher education, the ground is shaking due to artificial intelligence, which the representative, who, in addition to her parliamentary duties, also teaches courses at the Faculty of Philosophy in Zagreb, must be aware of.
At this moment, there are many students around the world, and perhaps in Croatia as well, who are idly scrolling through Instagram while ChatGPT independently writes essays and homework for them. For now, many schools around the world are trying to resist these trends by banning ChatGPT in classrooms, while older generations like the progressive Noam Chomsky grumble and say that artificial intelligence is high-tech plagiarism which avoids the learning process. Of course, there are also those who, like Harvard researcher Chris Dede, recommend that we shouldn’t resist computers but that they should evolve into new ways of education, and by the way, calculations about how many teachers and professors will lose their jobs appear in the media.
Algocratic regime
And finally, artificial intelligence may come for our good parliamentarians and replace them with algorithmic automation. The idea of an algocracy, rule by AI algorithms, which, among other things, would be in continuous contact with citizens through applications on smart devices and make decisions quickly with a significantly smaller number of officials, is not completely new – in Denmark, for example, experiments have already been made with computers that decide on local social measures, and the emergence of ChatGPT and similar tools will only further encourage supporters of such a management method. A Spanish survey showed that over 50% of Europeans would happily replace their representatives with an algorithm, with the percentage rising to 60% among younger people. According to Italian-American research, the more dissatisfied the citizens are with the people they elected, the more willing they are to leave the rule to a computer.
The initial costs of installing computer equipment for such a social arrangement are significant, but in the long run, a British study concludes, a bureaucracy led by artificial intelligence can be more efficient and transparent. Algocracy, however, can also turn into a monster if a very small number of ill-intentioned but capable people find ways to control the algorithms and turn the whole system into something like a digital monarchy. Some more conscientious politicians around the world took all these challenges more seriously, such as the 72-year-old representative from the US Congress, Don Beyer, who decided to enrol in a master’s degree programme in the field of advanced computer learning methods and digital neural networks. “The explosion of the availability of all knowledge to everyone on the planet will be a very good thing – and a very dangerous thing”, says Beyer, who only gets to attend one course a semester due to his many commitments in Congress.
Just water under the bridge
It is precisely the US Congress, which is expected in vain to legislate high technologies, that chronically suffers from technological illiteracy and is now trying to engage experts and advisers who will explain to the representatives what it is all about. American and European political leaders’ trotting for digital innovations has been going on for decades and will further deepen due to the acceleration that awaits us, which could lead to social crises.
ChatGPT is, namely, only a small segment of the application of artificially intelligent technologies, which, according to ambitious announcements, can become a million times faster than today within just ten years.
The concern about whether artificial intelligence worships Tudjman or Tito more, expressed by the Club of Representatives of Most, could therefore soon be just an insignificant drop in the technological flood that is encroaching and threatening to reach us all. Then we may only be left to wonder why we didn’t react more timely and hum the lines from the recent country hit “Love was fun and life was easy, now it’s just water under the bridge”.