The newest iteration of OpenAI’s generative chatbot, GPT-4, is heightening existential fears over misinformation, disinformation and job security. Its reasoning abilities are purportedly better than ChatGPT, with OpenAI pointing to the newest version’s superior scores on standardized tests like the LSAT, bar exam and GRE as evidence. GPT-4 can understand the semantic content of images and can accurately discern why a meme is funny. It is capable of translating hand-drawn pictures of web pages into fully functional ones, instantly.

Given only the image, GPT-4 explains why a global map made of chicken nuggets is funny. (Screenshot/GPT-4/@inthesun_x)

Prominent figures in Silicon Valley — including Elon Musk, Apple co-founder Steve Wozniak and former presidential candidate Andrew Yang — just signed a petition to halt developments on artificial intelligence stronger than GPT-4. The petition has over 30,000 signatories and cites large-scale societal risks as its founding purpose.

“Systems with human-competitive intelligence can pose profound risks to society and humanity, as shown by extensive research and acknowledged by top AI labs,” the letter reads.

In an interview with ABC News, OpenAI founder Sam Altman expressed his fears over AI’s potential for spreading disinformation.

“I’m particularly worried that these models could be used for large-scale disinformation,” Altman said in a March interview with ABC News’ Rebecca Jarvis. “Now that they’re getting better at writing computer code, (they) could be used for offensive cyberattacks.”

But Altman dispelled fears of AI world domination, saying “this is a tool that is very much in human control.”

Researchers and fact-checkers working in artificial intelligence have warned of spurious statements surrounding the capabilities of chatbots.

“Claims that generative AI tools will produce accurate content should be treated with great caution. There is nothing inherent in the technology that provides any assurance of accuracy,” said Kate Wilkinson, senior product manager at Full Fact, a London-based fact-checking organization that implements AI. “ At this stage it is too early to be able to say exactly how this will evolve, but it is clearly a significant change in the general accessibility of such a powerful set of tools to such a wide group of users.”

Others echoed Altman’s worries over mass dissemination of disinformation.

“A cunning manipulator could use it to quickly generate a lot of information that sounds trustworthy, but is, in fact, misleading or utterly false,” said Marcel Kiełtyka, a spokesperson for Demagog, a nongovernmental and fact-checking organization based in Poland. “Fact-checkers should keep an eye on developments of such technologies because the automation of content-creation poses a risk of flooding social media with misinformation.”

Interesting fact-checks

(Shutterstock)

• 211 Check: Did Pope Francis say celibacy can be revised? (English)
  • “When asked by Infobae whether celibacy could be revised, Francis answered that it could. He added that celibacy in the Western Church is a ‘temporary prescription. … It is not eternal like priestly ordination.’”
• News Checker: Viral police advisory against perfume scam in Hyderabad found to be fake (English)
  • “An advisory, purportedly from ‘KR Nagaraju, DCP (crime branch)’ and addressed to all ‘Hyderabadis,’ cautioning citizens against being persuaded by ‘perfume salesmen’ who test their products in the parking lot of malls, is being widely shared.” But no such advisory was issued.
• Re: Baltica: Does Facebook censor pro-Ukraine posts? (Latvian) 
  • “Many in Latvia believe so, however, Re:Baltica found no evidence that Facebook content moderators deliberately delete pro-Ukraine posts. But another problem was revealed — the content policy is designed in such a way that it is practically impossible to talk emotionally about the war.“
• FactCheck Vlaanderen: Less agricultural land needed for a plant-based diet (English)
  • “In an interview with De Tijd, food scientist Louise Fresco states that a plant-based diet ‘requires about a quarter more agricultural land.’ The figure is based on a comparison in the amount of arable land between a plant-based diet and a scenario with less animal consumption. But arable farming is only part of the total agricultural land.”

Quick hits

(Shutterstock)

From the news: 

• Ex-President Duterte says he’s willing to face consequences of his ‘war on drugs’: “Reacting to the decision of the International Criminal Court (ICC) junking the Philippine government’s appeal to suspend its investigation of killings linked to his administration’s “war on drugs,” former President Rodrigo R. Duterte said on Wednesday he was ready to “face the music.” (Antonio Colina, Minda News)
• Journalist detained by Russia pleads not guilty to spying charges: “American journalist Evan Gershkovich, arrested this Thursday in Yekaterinburg, Russia, has pleaded not guilty to espionage charges before a court in Moscow, advances Russian state news agency Tass. The court decided that The Wall Street Journal journalist is in preventive detention for almost two months, until May 29th.” (Filipa Mendes, Ines Pereira, Publico)
• Russian law that punishes disinformation with 15 years in prison serves as a warning for debate in Brazil “Intentionally sharing ‘fake content’ about the military can get you up to 15 years in prison in Russia. That’s what the new law passed by the Russian parliament two weeks ago establishes.” (Cristina Tardáguila, Lupa)

From/for the community: 

• Celebrate International Fact-Checking Day the week of April 2. The International Fact-Checking Network is featuring op-eds from signatory organizations across the globe. We’re also hosting two panel discussions on fact-checking during crisis and fact-checking organizations facing oppressive lawsuits across the world. Sign-up here!
• Google and YouTube are partnering with the International Fact-Checking Network to distribute a $13.2 million grant to the international fact-checking community. “The world needs fact-checking more than ever before. This partnership with Google and YouTube infuses financial support to global fact-checkers and is a step in the right direction,” said Baybars Örsek, former executive director of the IFCN. “And while there’s much work to be done, this partnership has sparked meaningful collaboration and an important step.”
• The IFCN has awarded $450,000 in grant support to organizations working to lessen the impact of false and misleading information on WhatsApp. In partnership with Meta, the Spread the Facts Grant Program gives verified fact-checking organizations resources to identify, flag and reduce the spread of misinformation that threatens more than 100 billion messages each day. The grant supports eleven projects from eight countries: India, Spain, Nigeria, Georgia, Bolivia, Italy, Indonesia and Jordan. Read more about the announcement here.
• IFCN job announcements: Program Officer and Monitoring & Evaluation Specialist

Thanks for reading. If you are a fact-checker and you’d like your work/projects/achievements highlighted in the next edition, send us an email at factually@poynter.org by next Tuesday. Corrections? Tips? We’d love to hear from you. Email us at factually@poynter.org.

Factually is a newsletter about fact-checking and misinformation from Poynter’s International Fact-Checking Network. Sign up here to receive it in your email every other Thursday.