ChatGPT blocked 250,000 picture generations of presidential candidates

OpenAI estimates that ChatGPT rejected greater than 250,000 requests to generate pictures of the 2024 U.S. presidential candidates within the lead as much as Election Day, the corporate mentioned in a weblog on Friday.

The rejections included image-generation requests involving President-elect Donald Trump, Vice President Kamala Harris, President Joe Biden, Minnesota Gov. Tim Walz and Vice President-elect JD Vance, OpenAI mentioned.

The rise of generative synthetic intelligence has led to considerations about how misinformation created utilizing the expertise might have an effect on the quite a few elections going down around the globe in 2024. 

The variety of deepfakes has elevated 900% yr over yr, in line with information from Readability, a machine studying agency. Some included movies that had been created or paid for by Russians in search of to disrupt the U.S. elections, U.S. intelligence officers say.

In a 54-page October report, OpenAI mentioned it had disrupted “greater than 20 operations and misleading networks from around the globe that tried to make use of our fashions.” The threats ranged from AI-generated web site articles to social media posts by pretend accounts, the corporate wrote. Not one of the election-related operations had been capable of entice “viral engagement,” the report famous.

In its Friday weblog, OpenAI mentioned it hadn’t seen any proof that covert operations aiming to affect the end result of the U.S. election utilizing the corporate’s merchandise had been capable of efficiently go viral or construct “sustained audiences.”

Lawmakers have been significantly involved about misinformation within the age of generative AI, which took off in late 2022 with the launch of ChatGPT. Massive language fashions are nonetheless new and routinely spit out inaccurate and unreliable data.

“Voters categorically mustn’t look to AI chatbots for details about voting or the election — there are far too many considerations about accuracy and completeness,” Alexandra Reeve Givens, CEO of the Heart for Democracy & Know-how, advised CNBC final week.

WATCH: AI more likely to be much less regulated and extra risky beneath second Trump presidency