Georgia AI Evidence Law: What 2026 Means for Claims

The Georgia General Assembly has thrown a wrench into how we handle evidence from AI-powered devices in personal injury claims. Starting July 1, 2026, House Bill 789 (HB 789) is going to force a whole new set of foundational rules on us, especially for data coming from AI glasses used in a medical context to document something like a motorcycle accident. This law completely changes the game for attorneys trying to use AI motorcycle data in court, creating some big new hurdles and a few opportunities for both sides.

Understanding House Bill 789: New Standards for AI Evidence

HB 789 is Georgia’s first real attempt to grapple with emerging tech in the courtroom. The bill gets right to the point, amending O.C.G.A. § 24-9-901 (the main rule for authenticating evidence) and O.C.G.A. § 24-9-902 (for self-authenticating evidence). The meat of the change is a new subsection, O.C.G.A. § 24-9-901(c), that lays out exactly how to authenticate data from AI systems, like the kind embedded in AI glasses being used by doctors and paramedics.

Until now, getting digital evidence admitted was a pretty straightforward affair, usually just requiring testimony from someone who knew where it came from. But AI is often a black box, and that complexity demanded a new, more structured approach. The legislature saw the writing on the wall: without clear rules, courts would be all over the map evaluating AI reliability, especially in big-money injury cases where medical records are everything. The point of this law is to set a high, clear bar for trustworthiness before some AI-generated report can sway a jury.

Who is Affected by These Changes?

HB 789 is going to affect a lot of people in Georgia’s legal and medical fields. Personal injury attorneys on both sides of the aisle are on the front lines. Plaintiff’s counsel now has a new set of hoops to jump through when using AI glasses data to prove the extent of a client’s injuries or show how care was provided. On the other side, defense counsel has a whole new playbook for attacking the admissibility of that evidence by digging into the AI’s methodology and data integrity. It levels the playing field, but it also means lawyers need to get a lot more technical, fast.

Healthcare providers are also in the hot seat, particularly those using AI glasses for things like patient exams or post-op checks. Any hospital, clinic, or practice in Georgia using this tech has to get its data management house in order. That means keeping careful logs of device use, calibration, and data exports. If they don’t, that critical patient data could get thrown out of court, which torpedoes the patient’s claim and could even open the provider up to liability. Imagine a trauma center in downtown Atlanta using AI glasses on a patient from a wreck at the I-75/I-85 interchange. Every single byte of data they collect has to meet these new rules or it’s gone.

And don’t forget the AI device manufacturers and developers. If they want to sell their tech to Georgia’s healthcare market, their products better be designed to produce evidence that can stand up in court. They’re going to have to provide exhaustive documentation on their AI models, training data, validation methods, known biases, to help the doctors and lawyers who have to defend this stuff on the stand.

Key Procedural Changes for Admitting AI Evidence

The single biggest change from HB 789 is the pre-trial notice requirement. Under the new O.C.G.A. § 24-9-901(c)(1), if you plan to introduce evidence from an AI system, you must give every other party written notice at least 60 days before trial. Take this notice seriously. It has to include:

• The specifics of the AI system: manufacturer, model, and version.
• Information on the actual algorithm that generated or analyzed the data.
• All documentation you have on the system’s calibration, maintenance, known error rates, or other limitations.
• A straight-up explanation of how the data was collected, processed, and stored.
• The exact data points or outputs you want to show the jury.

This long notice window gives the other side time to dig in, hire their own experts, and get their Daubert challenges ready. On top of that, O.C.G.A. § 24-9-901(c)(2) says the person offering the AI evidence has to prove (by a preponderance of the evidence) that the system was working correctly when the data was made and that nobody has messed with it since. This usually means you’ll need live testimony from a qualified witness, a tech, a data scientist, or a doctor who actually knows the device. You can’t just say “it came from an AI” anymore.

For example, a plaintiff’s attorney wanting to use footage from a paramedic’s AI glasses at a motorcycle accident on Peachtree Road has to file that detailed notice. It would have to specify something like “MedVision AI Goggles, Model MV-300, firmware v2.1,” name the object recognition and vital sign algorithms used, and include the last calibration log from the manufacturer, MedTech Solutions, Inc.

Establishing Foundational Reliability: The Expert Witness Role

These amendments make expert witnesses more important than ever in cases with AI evidence. While the law doesn’t force you to use an expert every single time, good luck getting complex AI data admitted without one. You’ll need to satisfy O.C.G.A. § 24-7-702, Georgia’s version of the Daubert standard, by having an expert show that:

1. Their specialized knowledge will actually help the jury understand the evidence.
2. The testimony is built on solid facts or data.
3. The testimony comes from reliable principles and methods.
4. They applied those methods and principles correctly to the facts of your case.

When you’re talking about AI glasses in medicine, your expert could be a biomedical engineer who can unpack the physiological monitoring algorithms. Or maybe it’s a computer scientist who can vouch for the AI model’s accuracy, or a forensic analyst who can confirm the data’s integrity. Without that kind of validation, especially for a newer AI application, a judge in Fulton County or Gwinnett County Superior Court is very likely to decide the evidence is just too unreliable to show a jury.

Get your experts involved early. Seriously. Trying to find someone on the eve of trial who can take apart a complex AI system is a great way to lose a motion. The guts of these algorithms and their potential for mistakes aren’t obvious. A good expert can explain these issues clearly to a judge and jury, which can make or break the credibility of your AI evidence, or give you the ammo you need to tear down your opponent’s.

Practical Steps for Attorneys and Healthcare Providers

For Attorneys:

1. Review Existing Cases: Go through your active PI files right now and flag any that might involve AI-generated data from things like AI glasses.
2. Update Discovery Protocols: Your standard requests for production need to be updated to ask specifically about the use of AI systems. You need to demand all the documentation on any AI evidence the other side might have.
3. Educate Your People: Make sure your team and your clients know about these new rules. If your clients are doctors or hospitals, they need to hear from you about how important their record-keeping is now.
4. Line Up Experts Early: If AI is in your case, start talking to tech experts now. You need someone who can assess the system’s reliability and get you ready for a Daubert fight.
5. Prepare to Fight: Get ready to challenge the other side’s AI evidence. You can attack it for bad notice, improper authentication, or just because the AI itself is unreliable.

For Healthcare Providers:

1. Standardize AI Use: Write clear, internal policies for how your staff uses AI glasses and other smart medical devices. Cover everything from patient consent to data storage.
2. Keep Careful Logs: Every AI device needs a detailed logbook: calibration dates, maintenance, software updates, who was trained on it, and when. This is how you prove the device was working right.
3. Lock Down Your Data: You need secure systems to make sure AI data isn’t altered or deleted. Have a clear chain of custody from the moment the data is created until it’s archived.
4. Train Your Staff: Everyone who touches an AI device needs regular training on how to use it properly and how to handle the data so it holds up in court.
5. Talk to Your Lawyer: Stay in regular contact with your legal counsel to make sure your policies are keeping up with rules like HB 789 as AI tech continues to evolve.

Admitting digital evidence in Georgia just got a lot more complicated thanks to HB 789. For attorneys with cases involving healthcare data accidents, getting up to speed on these new requirements isn’t a choice if you want to advocate effectively. Winning these cases from now on will come down to preparing early, keeping perfect records, and bringing in the right experts.