Arizona lawmakers are hoping to make it easier for politicians to take legal action as the use of deepfakes increases during election campaigns.
House Bill 2394 could pave the way for candidates, and Arizonans at large, to challenge digital impersonations if they can prove the material was produced without consent and with the intent to spread disinformation.
Karthik Ramakrishnan is the founder and CEO of Armilla AI.
He said more focus should be placed on regulating artificial-intelligence models, that campaigns are using to ensure that mis- and disinformation don’t fall through the cracks.
“Take existing laws and extend them to cover the downsides risks of AI,” said Ramakrishnan, “which I think is very, very prudent – rather than creating brand new laws in place.”
The Federal Communications Commission recently banned robocalls that use voices generated by AI.
Ramakrishnan contended that AI isn’t bad technology in and of itself, and added that it’ll be hard to prevent AI-generated content from being produced – which is why he thinks Arizona’s proposed law is a step in a better direction.
Ramakrishnan contended that rather than solely focusing on the negatives, politicians should also use AI in their favor.
“If an official or a candidate wanted to leverage AI, have their voice generated in multiple languages when they call certain demographics, tailor their message to a specific constituency,” said Ramakrishnan. “Instead of recording 50 or 70, 100 different variations, they can have an AI generate those and speak to their potential voters.”
But data from Pew Research Center shows Americans are increasingly cautious about the growing role of AI in their lives – with 52% of Americans stating they’re more concerned than excited about AI in daily life.
For companies and organizations that use AI, Ramakrishnan encouraged them to willingly submit themselves to a third-party audit and post the results on their websites to instill more confidence and transparency.
For consumers of information in Arizona and around the country, he suggested having a stronger filter.
“We need to have that extra level of filter,” said Ramakrishnan. “Is the plausibility of that being true? And doing our own research. Not taking everything at face value, but doing that second or third order of research, because this is about voting, this is about democracy, it is about your franchise.”
Support for this reporting was provided by the Carnegie Corporation of New York.
