OpenAI CEO warns of artificial intelligence’s ability to persuade voters: It may 'cause significant harm to the world'

On Tuesday, OpenAI CEO Sam Altman appeared before Congress to address the need for new rules to help guide the development of artificial intelligence as the sector continues to grow exponentially.

He suggested that while he was committed to working with the government to ensure that humans remain in control of the technology, there was a chance that things could get out of control.


 

As the Daily Mail reports, Altman warned that if left unattended, AI could "go quite wrong," adding that OpenAI sought to "work with the government to prevent that from happening." He admitted that his "worst fear" was that products such as OpenAI's ChatGPT would cause "significant harm to the world."

During the hearing, Altman was confronted on a number of topics related to those fears. Senator Josh Hawley of Missouri, for example, questioned how the technology could be used to influence elections.

"Should we be concerned about models that can predict survey opinion and then help organizations and entities fine-tune strategies to elicit behaviors from voters?" he asked.

Altman called AI's ability to persuade people one of his "greatest areas of concern," noting that more work had to be done ahead of the 2024 election.

Sen. Richard Blumenthal of Connecticut suggested that the threat AI posed to certain jobs was the "biggest nightmare" facing the public, to which Altman responded that yes, some professions would eventually be automated out of existence.

He pointed out, however, that new jobs would be created as the technology progresses, though they will mostly be skilled and require extensive education.

In recent months, ChatGPT has taken the world by storm, emerging as both a godsend and a burden, depending on how one used it. The AI chatbot has been a boon to those seeking simple summaries of complex topics, however, it has also spewed outright lies and even generated false reports accusing an innocent professor of sexual assault.

Artificial intelligence is also responsible for "deepfakes," content manipulated to make it appear as though someone is saying or doing something. Deepfake videos of elected officials have polluted social media, with many believing what they are watching is actually real.