Follow us on Google News Follow Blog


OpenAI, Microsoft AI technologies make misleading election imagery, researchers warn

Smartphone is pictured in front of Microsoft logo displayed in this image shot July 26, 2021

March 6 (mod1s) - Image creation tools powered by artificial intelligence from companies including OpenAI and Microsoft (MSFT.O), opens new tab can be used to produce photos that could promote election or voting-related disinformation, despite each having policies against creating misleading content, researchers said in a report on Wednesday.
The Center for Countering Digital Hate (CCDH), a nonprofit that monitors online hate speech, used generative AI tools to create images of U.S. President Joe Biden laying in a hospital bed and election workers smashing voting machines, raising worries about falsehoods ahead of the U.S. presidential election in November.

"The potential for such AI-generated images to serve as 'photo evidence' could exacerbate the spread of false claims, posing a significant challenge to preserving the integrity of elections," CCDH researchers wrote in the paper.
CCDH examined OpenAI's ChatGPT Plus, Microsoft's Image Creator, Midjourney and Stability AI's DreamStudio, which can individually produce images from text prompts.
The allegation follows a statement last month that OpenAI, Microsoft and Stability AI were among a group of 20 tech firms that inked an agreement to work together to avoid false AI material from meddling with elections taking place internationally this year. Midjourney was not among the first set of signatories.

CCDH stated the AI tools created photographs in 41% of the researchers' tests and were particularly sensitive to prompts that called for photos suggesting electoral fraud, such as voting ballots in the garbage, rather than images of Biden or former U.S. President Donald Trump.
ChatGPT Plus and Image Creator were effective in blocking all suggestions when requested for photographs of applicants, says the research.
However, Midjourney fared the poorest out of all the tools, providing false visuals in 65% of the researchers' tests, it stated.

Some Midjourney photographs are accessible openly to other users, and CCDH claimed there is evidence some individuals are already exploiting the technology to produce deceptive political material. 

One effective prompt provided by a Midjourney user was "donald trump being arrested, good quality, paparazzi photo.”
In an email, Midjourney founder David Holz stated "updates related specifically to the upcoming U.S. election are coming soon," adding that pictures developed last year were not reflective of the research lab's current moderation methods.

A Stability AI spokesman said the business revised its standards on Friday to ban "fraud or the creation or promotion of disinformation."
An OpenAI spokesman claimed the business was attempting to avoid exploitation of its technologies, while Microsoft did not reply to request for comment.

Post a Comment

Cookie Consent
We serve cookies on this site to analyze traffic, remember your preferences, and optimize your experience.
It seems there is something wrong with your internet connection. Please connect to the internet and start browsing again.
AdBlock Detected!
We have detected that you are using adblocking plugin in your browser.
The revenue we earn by the advertisements is used to manage this website, we request you to whitelist our website in your adblocking plugin.
Site is Blocked
Sorry! This site is not available in your country.