TVPINFO — Wykrywacz klamstw
20260914 154200 UTC
305 transcript segments
Google Cloud Speech-to-Text API (Chirp) + Gemini 2.5 Flash Non-Thinking
Data courtesy of The GDELT Project (https://www.gdeltproject.org/), from the Internet Archive TV News Archive.
Machine transcription. Treat it as a searchable index of what was broadcast, not a verbatim quotation record.

[00:00:00] it's probably still slippery. Monika Andrzejczak,
[00:00:02] thank you. And now that's all, but we will have
[00:00:06] more topics for you at 6 PM in the main edition of Panorama. And before
[00:00:10] us now is the lie detector.
[00:00:22] Is the government experimenting on Polish children? Truth
[00:00:26] or lie? What is healthier?
[00:00:38] A politician cannot, well, it is known that it depends on the benefits that
[00:00:42] he can achieve by promoting either one or
[00:00:45] the other. We start the lie detector, the only daily program on
[00:00:49] Polish television, in which we unmask lies and
[00:00:52] manipulations. Magdalena Pernet,
[00:00:54] welcome.
[00:01:06] The leadership of the Ministry of Education introduces ideological trends to schools
[00:01:09] and experiments on our
[00:01:12] children. Such an alarm is raised by representatives of the National Movement,
[00:01:16] Krzysztof Bosak's people. On the screens you can see
[00:01:19] a graphic that they shared
[00:01:23] on Platform X. It's about, attention, the
[00:01:27] planetary diet, about the fact that...
[00:01:30] from September it was introduced to school
[00:01:33] canteens. In reality, despite the suggestions
[00:01:37] of politicians of the National Movement. The diet is, the planetary diet
[00:01:40] is not an ideological invention
[00:01:44] of the ruling party, it is a nutrition model developed by
[00:01:47] scientists, experts in the field
[00:01:50] of dietetics. This is not a new
[00:01:53] topic, this is a topic that was published in
[00:01:57] a great journal, The Lancet. in
[00:02:00] 2019, as I mentioned, and it was
[00:02:03] created by 37 specialists from 16
[00:02:07] countries, that is, experts in the field of food and
[00:02:11] nutrition and dietetics.
[00:02:14] The planetary diet is based primarily on vegetables,
[00:02:17] fruits, legumes and nuts. It does not
[00:02:21] , however, assume a complete resignation from meat, fish, eggs or
[00:02:25] dairy, but a reduction in the consumption
[00:02:28] of these products. and this is what most outraged opposition
[00:02:32] politicians, among them Patryk Jaki,
[00:02:35] MEP of Law and Justice.
[00:02:38] I would like to brag to you that I practiced the planetary diet today,
[00:02:42] namely, first I ate a pork chop from planet Earth, then I jumped on
[00:02:46] my bike and went for another pork chop, also from planet Earth. I recommend such a
[00:02:49] planetary diet to everyone.
[00:02:53] Patryk Jaki, Doctor from the Faculty
[00:02:55] of National Security of the Academy of War
[00:02:58] Art. ridicules dietary recommendations
[00:03:02] developed by professors of dietetics, well, like
[00:03:06] that we won't get far.
[00:03:09] In the planetary diet, attention is paid precisely to ensuring
[00:03:13] that there are more products that are less processed, but
[00:03:16] also that there are more plant-based
[00:03:19] products, and we also know from various studies, these are
[00:03:22] recent publications, but also earlier ones, precisely using the
[00:03:26] planetary diet, that people who base their diet on
[00:03:30] a plant-based diet, who eat this way, live longer, and
[00:03:34] also get sick less often with various
[00:03:37] diet-related diseases, civilization diseases, but also this
[00:03:40] diet, ladies and gentlemen, allows for a greater degree of maintaining
[00:03:44] a healthy body weight, and as for our Polish
[00:03:48] children, there is also a lot to
[00:03:50] be done here. Patryk Jaki, instead
[00:03:54] of listening to experts and discussing on a decent level,
[00:03:57] prefers to do it just to spite, it's not clear who, and
[00:04:01] eat two cutlets, two dinners, and
[00:04:04] pork cutlets specifically
[00:04:07] in one day. This is, ladies and gentlemen, a serious politician,
[00:04:11] one of the leading representatives of the largest opposition
[00:04:14] party. Patryk Jaki is a hypocrite.
[00:04:18] Back in 2018, the politician promoted
[00:04:21] vegetarian dishes and you can see
[00:04:24] the evidence right now on your screens. At that time, Patryk Jaki was
[00:04:28] a candidate for the office of Mayor
[00:04:30] of Warsaw.
[00:04:34] For some, a sunny day is fuel for action, for
[00:04:38] others, morning coffee. But
[00:04:41] for Confederation politicians, unfortunately,
[00:04:44] Ukrainians who are trying to settle down in
[00:04:47] Poland are enough. Paweł Usiądek took on the issue of education
[00:04:51] for Ukrainian children and the fact that the Ukrainian
[00:04:54] minority establishes its own educational institutions
[00:04:57] in our country.
[00:05:00] Ukrainians are opening another school in Poland. A parallel society
[00:05:04] is emerging, which is the beginning of future problems.
[00:05:08] Let's compare this with the numbers. In Polish schools,
[00:05:11] 246
[00:06:08] UPA, Bandera, and other scoundrels from
[00:06:10] the East.
[00:06:14] It probably won't surprise you, it won't surprise you, but
[00:06:17] I'll say it clearly: Mr. Usiądek manipulates the facts,
[00:06:21] the Poznań school he mentioned
[00:06:23] is not a Ukrainian school, but
[00:06:27] a Ukrainian-Polish one, which means that not
[00:06:31] only is Ukrainian culture and identity nurtured in it
[00:06:34] for students, but also conditions are created for their integration with
[00:06:38] Polish society. It is also not, at least for now, a facility
[00:06:42] It is that provides
[00:06:45] full-time education for children enrolled in it.
[00:06:49] a form of additional activity,
[00:06:52] supplementing classes in their standard
[00:06:55] school. We are currently in the process of creating
[00:06:59] and developing our areas of activity. The basis of our
[00:07:02] activity are creative classes and workshops on various
[00:07:05] topics, aimed at developing children, their abilities, and
[00:07:08] interests. We also cooperate with other foundations and organizations.
[00:07:20] They have the opportunity to use additional educational
[00:07:23] support, including individual classes and tutoring, as
[00:07:27] well as participating in art therapy classes and receiving psychological
[00:07:31] support.
[00:07:33] Even if it were a school providing full-time
[00:07:36] education, national and ethnic minorities in
[00:07:40] Poland have the right to... create and
[00:07:43] run such institutions, it's not a matter of recognition or
[00:07:47] goodwill, such a right is guaranteed by Article
[00:07:50] 35 of the Polish Constitution. In such
[00:07:53] schools, representatives of a given minority have full
[00:07:57] influence over the curriculum, the language of
[00:08:00] instruction, and the teaching staff,
[00:08:02] but they must simultaneously meet the general requirements of Polish
[00:08:06] education law. This works analogously
[00:08:09] in, for example, Great Britain, doesn't it? and
[00:08:13] Polish children who live there benefit
[00:08:16] from it.
[00:08:19] OpenAI announces that its artificial intelligence model
[00:08:23] has solved one of the seven Millennium
[00:08:26] Problems, which means one of the greatest challenges
[00:08:29] of contemporary mathematics. I repeat, artificial
[00:08:33] intelligence was supposed to solve a problem that
[00:08:36] scientists had not been able to solve for
[00:08:40] decades. It sounds sensational, but
[00:08:43] at the same time terrifying. The verification
[00:08:46] of these reports is ongoing. What will change if OpenAI's
[00:08:49] theory is confirmed? Among other things, Witek Tabaka
[00:08:53] discussed this with Professor Piotr Sankowski, Director of the
[00:08:56] Idas Research Institute, at the Economic Forum
[00:09:00] in Karpacz. Hello, welcome from Karpacz,
[00:09:03] today my guest is Professor Piotr Sankowski, one of the leading
[00:09:06] Polish artificial intelligence researchers, a computer scientist from
[00:09:10] the University. Good day, Professor. Professor, let's start with this, with this
[00:09:13] information from the last few hours. OpenAI informs
[00:09:17] that with the help of artificial intelligence, one of the seven
[00:09:20] Millennium Problems has been solved.
[00:09:24] So, in order for us to know what
[00:09:28] we are talking about, let's start with these problems. Who established
[00:09:31] them, invented
[00:09:34] them? Who decided that for their solution there would be a million
[00:09:38] zlotys in prize money? And can we actually talk about
[00:09:42] some kind of solution here? That is, indeed, these are
[00:09:45] very important problems in mathematics or theoretical
[00:09:49] computer science, which were posed a few years ago, at the end
[00:09:52] of the millennium, when these billion-dollar, million-dollar
[00:09:56] prizes were designated, a million dollars for solving this
[00:10:00] problem. So far, it seems that one problem was solved
[00:10:03] by one of the mathematicians, now there is another one, where
[00:10:07] the statement that it was possible to solve it
[00:10:09] is put forward by OpenAI, which
[00:10:12] published a fairly long and complex paper
[00:10:16] on this topic, which is currently being verified by
[00:10:20] various people. Well, it seems to look very sensible
[00:10:23] indeed.
[00:10:33] differential equations, which have a very
[00:10:37] close influence on applications, i.e.
[00:10:40] physics, and it had been open for almost 100
[00:10:42] years. Mathematicians and physicists
[00:10:46] tried to solve this problem, so far unsuccessfully, but I have
[00:10:49] the impression that in trade media, let's say, in business media, for example
[00:10:53] on LinkedIn, we can read about both enthusiastic optimists who are
[00:10:57] happy that this great breakthrough has actually happened, and those who say
[00:11:01] that we are still very far from it. When will we know if
[00:11:04] this breakthrough has actually occurred or not? But a breakthrough with this problem
[00:11:08] or generally with this problem and whether with the
[00:11:10] help of OpenAI tools, this problem
[00:11:14] has been solved, well, we certainly have to wait, this verification
[00:11:17] always takes some time, but what
[00:11:20] we have been observing in mathematics for half a year or a year,
[00:11:24] well, also a very strong trend of formalizing proofs,
[00:11:28] for example in such a formal language as Lean, where
[00:11:31] indeed mathematical proofs, such as we normally see
[00:11:34] on a sheet of paper, can be formalized in such a
[00:11:38] programming language, one could say,
[00:11:42] which, and what is interesting, it is possible
[00:11:45] to compile this proof only when
[00:11:48] it is correct, so if such a proof can be compiled,
[00:11:51] well, that means its correctness. So,
[00:11:55] we are moving from verifying such
[00:11:58] proofs by mathematicians, who read these proofs,
[00:12:02] disagree, say something is missing, correct them, to
[00:12:05] such a model of mathematics, where at the same time, not yet in
[00:12:09] this case, but there are earlier cases where it was possible
[00:12:13] to solve some open problems, where at the same time or
[00:12:16] not much later, a formal proof also appears
[00:12:19] in the language of formal
[00:12:21] proof verification. So a computer is
[00:12:25] able to verify that the proof is correct. What we are observing right now
[00:12:28] is certainly a kind of race between companies dealing with artificial
[00:12:32] intelligence. To what extent, in your opinion, are these announcements that have recently appeared, not
[00:12:36] only from OpenAI, but many companies,
[00:12:40] a marketing race, to what extent can we actually talk about
[00:12:43] further breakthroughs in this field? I mean, we
[00:12:46] need to be aware that these future
[00:12:49] applications of this artificial intelligence or language models, well, they
[00:12:52] require such... I don't know,
[00:12:55] logical thinking skills, right,
[00:12:58] planning, performing complex operations
[00:13:02] or various complex reasoning, right? Well, if
[00:13:05] we run a company, then we need to plan a
[00:13:08] strategy, then execute it, it's like multi-month planning,
[00:13:12] or even if we run a research institute, right? Well, these aspects
[00:13:16] related to planning or logical planning of certain
[00:13:19] steps that follow each other, some of which may not succeed, so
[00:13:23] we also need to have some emergency
[00:13:26] plan of action,
[00:13:29] right? Well, the most basic test for doing this is
[00:13:33] mathematics, and that's what we see, right? This is
[00:13:36] this test field, such
[00:13:39] a testing ground for these models, which shows,
[00:13:43] proves, that well, these models
[00:13:46] understand this logic, they can plan, right,
[00:13:49] and they show that in the next steps,
[00:13:53] you can also engage models by, of course, including
[00:13:57] various data, right, because they can't do that yet. Mathematics is the easiest
[00:14:00] place to test various
[00:14:04] reasonings, right, because you can, for example, write this proof in
[00:14:07] formal language and automatically verify it.
[00:14:11] This shows that these applications
[00:14:14] will be coming, right, because these models
[00:14:17] can plan. OpenAI also presented the latest
[00:14:21] Astra model,
[00:14:24] allegedly already
[00:14:26] close to artificial general intelligence,
[00:14:30] right? So we can say that
[00:14:32] it's either at the level of the human mind or even
[00:14:36] sometimes better than the human mind.
[00:14:39] Do you think we have already reached this point? That is,
[00:14:42] in many aspects, these models have been
[00:14:46] surpassing humans for years
[00:14:49] in either average people, that is, well,
[00:14:52] of course, we use the concept
[00:14:55] of GI or other various benchmarks
[00:14:59] that we don't understand, but a concept that
[00:15:02] is much more measurable is simply putting
[00:15:06] these models to intelligence tests, that is, measuring their
[00:15:10] intelligence quotient, so we are not yet surpassing an outstanding person,
[00:15:13] so these models for
[00:15:16] a few years or, I don't know, half a year, it's not known how it will
[00:15:20] be, well, they achieve roughly 140, 150 IQ
[00:15:23] points, so
[00:15:29] because it's 100, so they operate at a level
[00:15:32] of solving logic puzzles, well, probably
[00:15:36] better than. 95% of people, right?
[00:15:39] So, of course, we will achieve
[00:15:42] AGI when this artificial intelligence is better than us in
[00:15:45] everything, better than all people, or when it's better than
[00:15:49] 99% of people, right? We are indeed very
[00:15:52] close to that moment where it is
[00:15:56] simply comparable to the best
[00:15:58] experts in any field
[00:16:02] of knowledge or science. We will stop here. Professor Piotr
[00:16:05] Sankowski, thank you very much for this conversation. That's all for now from
[00:16:08] Karpacz.
[00:16:12] At the end of the lie detector, we return to the story of Kamil
[00:16:15] Kowalenko, an alleged councilman from Płock,
[00:16:19] an alleged politician of the Civic Coalition,
[00:16:22] of Ukrainian origin. This man does not actually
[00:16:26] exist. Nothing matches, neither the name, nor
[00:16:29] the function, nor the party affiliation. Someone
[00:16:32] used the image of a real person, used artificial
[00:16:36] intelligence and created a fake
[00:16:39] profile on the internet. Many people took this story
[00:16:43] seriously, which is why we addressed it before
[00:16:45] the weekend on the lie detector. This
[00:16:49] is this gentleman, or rather, no, not a gentleman, a fictitious
[00:16:52] character, generated. How
[00:16:55] to create such a character, is it super complicated?
[00:16:58] Exactly, a few years ago it was complicated, because you needed tools
[00:17:02] like Photoshop and you had to use the tools skilfully,
[00:17:05] you had to have
[00:17:07] knowledge. Nowadays it is enough to enter
[00:17:11] the appropriate command, i.e. a prompt, and it
[00:17:15] is practically free and very fast, and
[00:17:18] unfortunately easy. That is why generating such a person is a matter of
[00:17:22] moments. Some internet users
[00:17:25] criticized us for this verification, and an example
[00:17:29] of a critical post against us can be seen
[00:17:32] now on your screens. Accusations appeared that
[00:17:35] public television should not deal with such problems,
[00:17:39] such topics, that it is not appropriate, that we are dealing with
[00:17:42] unserious matters, but is this really
[00:17:46] a topic that is not suitable for the lie
[00:17:49] detector? We dealt with this matter because many
[00:17:53] people believed the story of the fake
[00:17:57] Ukrainian councilman of the Civic Coalition from
[00:18:00] Płock.
[00:18:02] Why isn't he on the front? He's a Bandera-face, send him to the front,
[00:18:06] right? The voters' ignorance, wake
[00:18:09] up, Poles, my grandfather told me
[00:18:12] that there would come a day when Ukrainians and Jews would emerge from
[00:18:16] the North and want to make Poland their country, their laws, and that is
[00:18:20] happening. Mr. Kalach,
[00:18:23] a user who ridiculed us for
[00:18:26] raising this topic, in another of his
[00:18:30] posts, himself noted that many people
[00:18:33] believed in the fake councilman, just
[00:18:36] an hour before.
[00:18:50] Mr. Kalach, what exactly is your point? Are you criticizing us just for the sake of it, because it's TVP, so
[00:18:53] you have to?