AI-generated tweets considered more trustworthy than humans': Study
4 Min Read
A study shows that people may be more susceptible to false information if it is generated by AI.
-
The Ares, a humanoid bipedal robot designed by Chinese college students with funding from a Shanghai investment company, is displayed during the World Robot Conference in Beijing, on October 21, 2016. (AP)
People were found to be more inclined to trust content produced by OpenAI's GPT-3 model in a recent study that contrasted tweets produced by humans with those produced by AI.
Unexpectedly, the test's findings showed that, regardless of how accurate the information was, people preferred tweets written by AI because they seemed more credible.
People trust AI-generated tweets than those made by humans
After testing its participants, a study published in the Science Advances journal asserts that artificial intelligence (AI) produces reliable and engaging social media content more effectively than humans.
Participants found it challenging to distinguish between tweets created by real Twitter users and those created by the AI tool when the researchers presented them with various tweets produced by humans and GPT-3.
They were also exposed to fake news and deceptive information written by an AI language model that focused on issues like vaccines and climate change, which are the targets of numerous online misinformation campaigns.
Read next: Mercedes joins AI race, integrates ChatGPT into 900,000 vehicles
However, The Verge reported that the participants could not tell whether these were true or false facts, and that the only way they could tell if the information was accurate was if it had been created by GPT-3 rather than a human.
The study's lead author, Giovanni Spitale, a postdoctoral researcher and data manager at the University of Zurich's Institute of Biomedical Ethics and History of Medicine, said, "These kinds of technologies, which are amazing, could easily be weaponized to generate storms of disinformation on any topic of your choice."
Hence, in order to conduct this study, the researchers used Facebook ads to solicit responses from 697 participants.
Is AI the Future of Information Dissemination?
One of the most important technological advancements of our time is artificial intelligence (AI), and many have embraced this technology to help the various industries it supports.
AI has been embraced by many businesses, including numerous news and media organizations, to assist employees. This study's assertion that AI is more persuasive and reliable for the audience, however, suggests that the future may change and that many people may receive inaccurate misinformation.
OpenAI, Microsoft, and Twitter
The most cutting-edge technology available today is artificial intelligence (AI), and Microsoft and OpenAI are leading the way with significant new developments while many other businesses scramble to keep up.
Recently, ChatGPT-4 from OpenAI was able to connect to the internet to expand its functionalities. It was made possible by the use of plugins that give ChatGPT access to websites and other online resources so that it can respond to user inquiries more pertinently and precisely.
Another study found that AI could be used to trick cybersecurity professionals into not performing their jobs accurately and correctly. Researchers from Georgetown University's study on AI misinformation found that it could be used to trick even professionals and academics.
Read next: Human extinction by AI ‘not that concerning… at least for now': Expert
The study has demonstrated how susceptible people are to being duped by actors who may be using AI to their advantage, even experts. This might lead to its application in the field of cybersecurity, deceiving people into lowering their guard or pressuring these professionals to play into the hands of hackers for potential access in the future.
Twitter and Meta are two social media companies that are currently stepping up their AI capabilities to enhance their services. Spitale pointed out that the best long-term approach to combating misinformation was to go "low-tech", promoting critical thinking abilities so that people would be better able to tell the difference between fact and fiction.
