President Donald Trump and Georgia lawmakers both say they want to encourage innovation in the field of artificial intelligence. But they are taking different approaches.

In Georgia, Republican legislators are pushing for regulations to rein in the technology.

Some proposals would outlaw AI-generated child pornography and election deepfakes. Other measures would require state agencies to disclose any use of AI and limit what kind of data companies can collect.

State lawmakers frame their proposals as promoting AI innovation while implementing light regulations around the technology to protect Georgians from misinformation and unwanted data collection. But civil liberties groups say those efforts don’t go far enough to protect consumers.

House Technology Committee Chairman Todd Jones, R-South Forsyth, said he sees AI as a new industrial revolution, and Georgia should be a leader in fostering that.

“We understand that great innovation requires a wide berth,” he said. But new technologies require “guardrails in terms of what you can’t do.”

While Georgia and other states are aiming for more regulation, Trump is seeking to roll some federal limits back. Within the first week of his return to the White House, he signed an executive order to rescind AI-related policies implemented during the Biden administration that placed some regulations on tech companies and sought to improve the AI readiness of federal agencies.

Trump has also called for the development of an AI Action Plan, which will outline policies to “enhance America’s position as an AI powerhouse and prevent unnecessarily burdensome requirements from hindering private sector innovation.”

A patchwork of state laws

AI policy experts view the president’s approach makes it more challenging for states grappling with the technology.

“If you have the federal government back off, it really just goes down to the states in terms of what regulation they want to do and how they want to do it,” said Adam Pah, assistant dean of digital innovation at Georgia State University. “And it’ll create a patchwork because many states will feel no impetus to do anything and other states will.”

That patchwork, Pah said, will likely create a difficult environment for some tech companies to navigate.

Many tech giants can absorb violations of state AI regulations and brush them off as a cost of doing business, but startups don’t have that luxury, Pah said.

“It’s very difficult for small companies to navigate this complexity,” he said.

State Rep. Brad Thomas, R-Holly Springs, is the primary author of House Bill 147 on artificial intelligence. (Arvin Temkar / AJC)

Credit: arvin.temkar@ajc.com

icon to expand image

Credit: arvin.temkar@ajc.com

The breadth of AI use across state agencies is unclear, but under House Bill 147, sponsored by Rep. Brad Thomas, R-Holly Springs, agencies would be required to report any use of AI to the Georgia Technology Authority, disclosing by year’s end the name of the AI tool, its capabilities and whether the system is being used without humans. Local governments would be required to do the same by the end of 2027.

“We feel like it’s important to get roots of where we are at. That way we can set policy about where we want to go,” Thomas said.

The legislation would require the Georgia Technology Authority to develop a set of model AI policies and procedures by 2026.

Senate Bill 111, sponsored by Sen. John Albers, R-Roswell, would require large companies to allow consumers to opt out having their personal data used for things such as targeted advertising. Consumers could also request companies delete their data.

In order for Georgia to be the best place to do business, Thomas said it needs to be the best place for AI. To do that, there needs to be some regulation in place, he said.

Opponents of the bill, such as the American Civil Liberties Union of Georgia, say that it doesn’t do enough to protect consumers, because it requires them to opt out of companies using their private data rather than opt in.

For deepfakes, lawmakers are taking it a step further.

A morphed version of Senate Bill 9 would do that by outlawing political deepfakes — images, videos or robocalls that mirror the likeness or voice of political candidates.

Jones said that voters need to be able to trust that the information they are receiving is accurate.

For example, he pointed to the use of a deepfake that mimicked the voice of then-President Joe Biden last year. The deceptive robocall discouraged voters from voting for Biden in the New Hampshire primary.

SB 9, also sponsored by Albers, would make it a misdemeanor to broadcast or publish deceptive information within 90 days of an election with the intent to influence a candidate’s chance of being elected, create confusion about election administration, or influence the result. For repeat offenders, it would become a felony.

Another House proposal, nearly identical to the original language of SB 9, would also criminalize the distribution of sexually explicit, AI-generated content involving children.

The AI proposals have all passed at one chamber of the Legislature and await action in the other as Sine Die — the last day of the legislative session — approaches on April 4.

About the Author

Featured

Georgia Power's Plant Bowen in Cartersville is shown in this 2015 photo. (Hyosub Shin/Atlanta Journal-Constitution/TNS)

Credit: hshin@ajc.com