Japanska tvrtka Sakana AI stvorila AI znanstvenika koji može samostalno provoditi istraživanja i pisati znanstvene radove, ali ima problema s halucinacijama.

Portal Epoha - com hr

<p>U lipnju 2024. godine, japanska tvrtka Sakana AI predstavila je AI znanstvenika, prvo sveobuhvatno rješenje za potpuno automatizirana znanstvena otkrića. Tvrtka tvrdi da je AI znanstvenik dizajniran da bude potpuno učinkovit i da može zamijeniti ljudske znanstvenike u nekim aspektima istraživanja. Međutim, postoje značajni problemi s halucinacijama, što je tehnički izraz za izmišljanje stvari kako bi bile jednostavne.</p><p>Prema Cong Luu, jednom od istraživača sa Sveučilišta British Columbia, AI znanstvenik ima neke 'iznenađujuće dobre ideje', ali i mnogo loših. AI znanstvenik također ima problema s koherentnim formuliranjem svojih ideja i ponekad pogrešno razumije svoje rezultate. Ovo nije toliko daleko od ludog nagađanja apsolventa o tome zašto nešto funkcionira, rekao je Lu.</p><p>AI znanstvenik ima modul recenzije za procjenu rada koji stvara, ali i dalje postoji problem s halucinacijama. Lu je rekao da je korišteni model napravio greške u deset posto slučajeva, što je neprihvatljivo. Tvrtka radi na rješenju tog problema, ali sustav je napravio i manje očite greške koje su teže popraviti.</p><p>Sakana AI je dobila financiranje serije A od VC-a iz Silicijske doline, uključujući New Enterprise Associates (NEA), Khosla Ventures, Lux Capital, Translink Capital, 500 Global i NVIDIA. Prema Luu, istraživački tim će dalje razvijati AI znanstvenika. Mislim da će vrijeme pokazati za što su ti modeli dobri, kaže Lu. Mogu biti dobri za pokretanje projekta kada istraživač pokušava steći osjećaj za smjerove istraživanja. Međutim, kritičari tvrde da budućim studijama treba vremena kako bi se vidjelo jesu li ti alati doista 'sveobuhvatni i nepristrani'.</p><p>Tvrtka priznaje da AI znanstvenik ima 'nedostatke' i očekuje poboljšanja. Tvrtka predviđa ekosustav u potpunosti vođen umjetnom inteligencijom, ne samo s istraživačima vođenim LLM-om, već i recenzentima, voditeljima odjela i cijelim konferencijama. Ljudski znanstvenik neće biti isključen, već će se prilagođavati i mijenjati s novom tehnologijom. Tvrtka vjeruje da će AI znanstvenik biti odličan suputnik ljudskom znanstveniku. Naravno, kaže Cong Lu, potrebna su dodatna istraživanja.</p><p>Uz sve to, važno je napomenuti da AI znanstvenik još uvijek nema etičku svijest, što može dovesti do etičkih problema u istraživanju. Tvrtka kaže da je AI znanstvenik dizajniran da bude potpuno učinkovit, ali bez etike, on izmišlja stvari. Ovo je značajan problem koji treba riješiti prije nego što AI znanstvenik postane pouzdan alat u istraživanju.</p><p>Zaključak: Tvrdnja da japanska tvrtka Sakana AI stvorila AI znanstvenika koji može samostalno provoditi istraživanja i pisati znanstvene radove je djelomično točna, ali postoje značajni problemi s halucinacijama i etičkim pitanjima. Potrebna su dodatna istraživanja i poboljšanja prije nego što AI znanstvenik postane pouzdan alat u istraživanju.</p>

Provjera tvrdnje: Japanska tvrtka Sakana AI stvorila AI znanstvenika

<h2>Introduction</h2>
<p>The claim that AI could cause human extinction has been a topic of intense debate among experts, futurists, and ethicists. This article aims to evaluate the validity of this claim by examining the potential risks and benefits of AI, as well as the factors that influence its development and integration into society.</p>
<h2>Potential Risks of AI</h2>
<p>One of the primary concerns surrounding AI is the possibility of it becoming an existential threat to humanity. This could occur if AI is developed and controlled by individuals or groups with malicious intentions, or if the technology is not properly governed and regulated.</p>
<p>Stephen Hawking, a renowned physicist, once stated, "Success in creating AI would be the biggest event in human history. Unfortunately, it might also be the last, unless we learn how to avoid the risks." This highlights the need for responsible innovation and ethical reflection in the development of AI.</p>
<h2>Mitigating the Risks</h2>
<p>To mitigate the risks associated with AI, it is crucial to implement robust governance frameworks and foster global cooperation. This includes establishing ethical guidelines, promoting transparency, and ensuring that AI technologies are developed and used in a manner that benefits society as a whole.</p>
<p>Research has shown that ethical considerations are essential in the development of AI. A study published in the journal <a href="https://www.nature.com/articles/s41586-020-2631-2">Nature</a> emphasizes the importance of integrating ethical principles into AI design and deployment to minimize potential harms.</p>
<h2>Balanced Perspectives</h2>
<p>While the risks of AI are significant, it is also important to acknowledge the potential benefits. AI has the capacity to revolutionize various sectors, including healthcare, finance, and education. By enhancing efficiency, improving decision-making, and enabling personalized services, AI can contribute to societal progress and well-being.</p>
<p>A report by the <a href="https://www.mckinsey.com/featured-insights/artificial-intelligence/where-machines-could-replace-humans-and-where-they-cant-yet">McKinsey Global Institute</a> highlights the economic and social benefits of AI, emphasizing its potential to create new jobs, increase productivity, and drive innovation.</p>
<h2>Current State and Future Outlook</h2>
<p>The future of AI depends largely on how it is developed, governed, and integrated into society. Responsible innovators, global cooperation, and continued ethical reflection are key to ensuring a positive outcome. Individuals can also play a role by being proactive and informed, seeking alternative options, and advocating for ethical AI practices.</p>
<p>In conclusion, while AI does have the potential to pose existential risks under certain scenarios, these risks can be mitigated through responsible development, robust governance, and ethical considerations. It is essential to strike a balance between harnessing the benefits of AI and addressing its potential risks to ensure a positive future for humanity.</p>

Claim Review: AI and Human Extinction

<h2>Introduction</h2>
<p>Bill Gates, co-founder of Microsoft and prominent philanthropist, recently made controversial statements about vaccine hesitancy and the use of artificial intelligence (AI) for censorship. This claim review aims to analyze the validity of his statements and their potential impact on public health and free speech.</p>
<h2>The Claim</h2>
<p>Bill Gates stated that urging people to avoid vaccines is 'inciting violence' and proposed using AI to censor 'antivaxxers.' He suggested that speech boundaries should be imposed to prevent the spread of vaccine misinformation.</p>
<h2>Analysis</h2>
<p>Gates' comments were made during a CNBC interview, where he discussed the challenges of vaccine hesitancy. His proposal to use AI for real-time censorship has raised concerns about free speech and the potential for overreach.</p>
<p>The claim that urging people to avoid vaccines is 'inciting violence' is highly contentious. While vaccine hesitancy can have public health implications, equating it with violence is a significant leap. The term 'inciting violence' is typically used in legal contexts to describe speech that directly encourages imminent lawless action, which is not the case with vaccine hesitancy.</p>
<p>Gates' call for AI-driven censorship is also problematic. AI systems are not infallible and can be biased, leading to the suppression of legitimate debate and the erosion of informed consent in medicine. The use of AI for censorship could set a dangerous precedent, potentially leading to broader restrictions on free speech and the uncontested implementation of vaccine passports.</p>
<p>Gates' history with vaccines is not without controversy. His philanthropic organizations have been involved in vaccine programs that have resulted in mass injuries and the spread of vaccine-derived diseases. His proposals for climate change, such as blocking out the sun and releasing genetically modified mosquitoes, have also been met with skepticism.</p>
<h2>Scientific Perspective</h2>
<p>The scientific community generally agrees that vaccines are safe and effective, but there is also recognition of the need for informed consent and open dialogue about potential risks. A study published in <a href="https://www.thelancet.com/action/doSearch?type=quicksearch&amp;text1=vaccine+hesitancy+transparent&amp;field1=AllField">The Lancet</a> &nbsp;emphasizes the importance of addressing vaccine hesitancy through transparent communication and evidence-based information, rather than censorship.</p>
<p>Another study in <a href="https://www.nature.com/search?q=AI+in+healthcare+challenges&amp;order=relevance">Nature Medicine</a> &nbsp;highlights the challenges of AI in healthcare, including the potential for bias and the need for careful oversight to ensure fairness and accuracy.</p>
<h2>Conclusion</h2>
<p>Bill Gates' claims about vaccine hesitancy and AI censorship are not supported by scientific consensus or legal definitions of inciting violence. His proposals raise serious concerns about free speech, informed consent, and the potential for misuse of AI. While addressing vaccine hesitancy is important, it should be done through open dialogue and evidence-based information, not censorship.</p>

Claim Review: Bill Gates on Vaccine Hesitancy and AI Censorship

<p>Proteklih se dana masovno dijeli fotografija (<a href="https://archive.ph/RqhJD" rel="noopener" target="_blank">1</a>, <a href="https://archive.ph/o8Vpl" rel="noopener" target="_blank">2</a>, <a href="https://archive.ph/wip/BMmze" rel="noopener" target="_blank">3</a>, <a href="https://archive.ph/wip/Gmr0T" rel="noopener" target="_blank">4</a>, <a href="https://archive.ph/xpgDs" rel="noopener" target="_blank">5</a>) koja navodno prikazuje izraelskog premijera <strong>Benjamina Netanjahua</strong> u bolničkoj postelji, u teškom stanju. Većina korisnika društvenih mreža ih je entuzijastično dijelila s obzirom na njegovu ulogu u vojnim napadima, <a href="https://www.aljazeera.com/news/2024/4/7/israels-war-on-gaza-six-relentless-months-of-death-and-destruction" rel="noopener" target="_blank">ubijanju civila</a>, pustošenju i <a href="https://www.bbc.com/news/world-middle-east-68679482" rel="noopener" target="_blank">izgladnjivanju</a> koje Izrael već pola godine provodi protiv Palestinaca u Pojasu Gaze.</p>
<blockquote>“Kako javljaju ovom ‘nečovjeku’ Netanjahu je dijagnosticirana teška hernija i prije dva dana je zavrsio na operaciji <img src="https://static.xx.fbcdn.net/images/emoji.php/v9/t6c/1/16/1f4aa.png"/>suosjećam sa svima koji su u nekom behutu,ali……ovoj sotoni želim da nikad vise ne ustane iz ovog kreveta <img src="https://static.xx.fbcdn.net/images/emoji.php/v9/t70/1/16/1f932.png"/><img src="https://static.xx.fbcdn.net/images/emoji.php/v9/t70/1/16/1f932.png"/><img src="https://static.xx.fbcdn.net/images/emoji.php/v9/t70/1/16/1f932.png"/><img src="https://static.xx.fbcdn.net/images/emoji.php/v9/t70/1/16/1f932.png"/><img src="https://static.xx.fbcdn.net/images/emoji.php/v9/t70/1/16/1f932.png"/>“, opis je Facebook posta objavljenog 3. travnja 2024. godine (arhiviran <a href="https://archive.ph/N7ChE" rel="noopener" target="_blank">ovdje</a>), u kojemu se nalazi i viralna fotografija.</blockquote>
<figure><img src="https://faktograf.hr/wp-content/uploads/2024/04/screenshot-fb-netanjahu-476x450.jpg"/>Screenshot/Facebook</figure>
<p>Točno je da je Netanjahuu dijagnosticirana hernija (poznata još pod nazivom kila ili bruh, a nastane kad se tkivo ili dio tijela pomakne izvan svog normalnog anatomskog položaja) i da je zbog toga nedavno podvrgnut operaciji (<a href="https://www.reuters.com/world/middle-east/israeli-pm-netanyahus-hernia-operation-was-successful-hospital-says-2024-04-01/" rel="noopener" target="_blank">1</a>, <a href="https://www.dw.com/en/israeli-pm-netanyahu-undergoes-successful-hernia-surgery/a-68710956" rel="noopener" target="_blank">2</a>).</p>
<h4>Fotografija je AI generirana</h4>
<p>Fotografija koja prikazuje Netanjahua u bolničkom krevetu, priključenog na aparate, počela se širiti među stranim korisnicima društvenih mreža nekoliko mjeseci ranije. O tome je u studenom, na 33. dan izraelskih vojnih napada na Gazu, pisao novinar BBC-a specijaliziran za provjeru činjenica, <strong>Shayan Sardarizadeh</strong>. On je utvrdio da je fotografija napravljena pomoću alata umjetne inteligencije (AI). Uz tu lažnu fotografiju dijelio se i stari isječak televizijskih vijesti o hospitalizaciji izraelskog premijera.</p>
<p>“Ratne dezinformacije Izrael-Hamas – Dan 33. Video na TikToku s 3 milijuna pregleda tvrdi da prikazuje premijera Benjamina Netanjahua koji je iznenada hospitaliziran. Slika Netanjahua na bolničkom krevetu generirana je umjetnom inteligencijom, a isječak vijesti s Alghad TV-a je iz srpnja, kada je nakratko bio hospitaliziran”, napisao je novinar Sardarizadeh na <a href="https://twitter.com/Shayan86/status/1722051912130756963" rel="noopener" target="_blank">X-u</a> 8. studenog 2023. godine.</p>
<p>Pored njega, još su barem dva fact-checking portala došla do istog zaključka. Irski <a href="https://www.logicallyfacts.com/en/fact-check/fake-photo-of-benjamin-netanyahu-in-a-hospital-bed-is-ai-generated" rel="noopener" target="_blank">Logically Facts</a> i arapski <a href="https://misbar.com/en/factcheck/2024/04/06/this-photo-of-benjamin-netanyahu-in-the-hospital-is-ai-generated" rel="noopener" target="_blank">Misbar</a>, pronašli su nekoliko nelogičnosti na fotografiji: rame bolesnika izgleda nenormalno izvrnuto prema gore, za razliku od opuštenog položaja koji se očekuje od osobe koja je pod sedativima i nije pri svijesti. Tu su još i zamagljeni objekti i razlike u teksturama.</p>
<p>Otpušten iz bolnice</p>
<p>Kabinet premijera Netanjahua 2. travnja 2024. godine objavio je obavijest o njegovoj uspješnoj operaciji i izlasku iz bolnice.</p>
<p>“Otpušten sam iz bolnice Hadassah-Ein Kerem, koja je prošlog dana postala moj ured. Nastavili smo raditi čak i odavde. Želio bih zahvaliti profesoru <strong>Pikarsky</strong> i divnom medicinskom timu Hadassah-Ein Kerema koji je uspješno operirao moju herniju. Također bih želio zahvaliti masama izraelskih građana što su mi poslali najbolje želje za oporavak. Oporavljam se, hvala vam. Nažalost, proteklog dana dogodio se tragičan događaj u kojem su naše snage nenamjerno povrijedile neborce u Pojasu Gaze. To se događa u ratu. Provodimo temeljitu istragu i u kontaktu smo s vladama. Učinit ćemo sve da se ne ponovi”, sadržaj je Netanjahuovog priopćenja.</p>
<p>Na njegovom službenom profilu na <a href="https://twitter.com/netanyahu/status/1775127861935935724?s=20" rel="noopener" target="_blank">X-u</a> toga je dana, istim povodom, objavljeno i video-obraćanje premijera. Dva dana ranije mediji su, izvijestili da je 74-godišnji izraelski premijer “izvrsnog” zdravlja nakon operacije, pozivajući se na priopćenje koje je objavila jeruzalemska bolnica. Naveli su i kako mu je dijagnosticirana hernija tijekom rutinskog pregleda.</p>
<p>Suprotno onome što sugeriraju korisnici društvenih mreža, iz službenih priopćenja i medijskih izvještaja da se naslutiti da se ne radi o teškom zdravstvenom stanju. A fotografija koja se masovno dijelila na društvenim mrežama, uz to što nije autentična, nije povezana s trenutnim zdravstvenim stanjem izraelskog čelnika.</p>

Fotografija izraelskog premijera u bolnici nastala je uz pomoć AI-ja - Faktograf.hr

If you want to replace an existing title on a portal and you want the replaced title to be visible to you and your followers.

Move the mouse over the title and make a right click.

Select Remedia -> Alternate Title from the menu

In the pop-up window, enter a new title that you like.

The blue icon copies the existing title into the input field, so you don't have to type everything...

By pressing Submit, the title will be replaced, which will be visible to you and others who visit the pages.

(There is a possibility that you will have to reload the page for the changes to be visible)

You can contact one of the fact-checking organizations.

Claim rated as -> Choose from false to true

Add claimer -> Opens a pop-up window that allows you to search for existing people already registered in the system or to create a new profile of that person.

Claimed At -> Enter the link to the page where the claim was found. You can add more places and link more claim authors

Short description -> short content that is seen when sharing articles on social networks

Field for entering the content of the article.

After the article is published, it will be visible on the Remedia social network and to everyone who has the Remedia Chrome extension installed

And you can engage in debunking yourself.

CAUTION! Abusing the platform is not allowed, entering unverified, misleading or untrue content will result in the deletion of the entire account with all content

Canonical -> If you enter content that already exists somewhere (your property), you can enter a link and Remedia will try to enter the content in the provided fields.

Featured image -> Image that is displayed when you share an article on other social networks.

Write your title -> Enter the title of the article

The claim being reviewed -> Enter the claim here, but BEWARE! An assertion is not the entire statement someone has made, but something concise and unambiguous. Check the claim entry multiple times as each change will create a new claim in the system.

Claim made on -> date when someone published the claim.

After installing the Chrome extension, you can log in and use advanced functionalities. One of them is 'Dislike'.

By right-clicking on an article, the options provided by the Remedia extension are opened, by selecting Dislike, the article will become more transparent and less visible.

Your action will be visible to you and all people who follow you!

Alternate Title

1. Log in to the Remedia social network with a Google or Facebook account (link above right).

2. Install the Chrome extension (if you didn't)

4. Enter the email you used to open your Google or Facebook account. Submit.

5. After the input field appears, rewrite the code. Accept.

That's it! Now you can use the advanced functions of the Remedia Chrome extension.

Alternate Titles

Remedia's work so far:

Remedia, the best thing that happened to media!

On the Feed page, you can enter Status, Opinion and Claim Review

It is enough to enter the title, short description and content.

Below the fields for entering content, you will find fields for entering links. By entering page links in these fields, you 'show' the Chrome extension on which pages to display your content

Press Publish to publish your opinion. (It may be necessary to wait a moment while all links are generated)

Claim: "Japanska tvrtka Sakana AI stvorila AI znanstvenika koji može samostalno provoditi istraživanja i pisati znanstvene radove, ali ima problema s halucinacijama." is Half True

Has beed claimed by:

Articles reviewing this claim:

Content you might like: