For elected members of county councils — the AI decisions that actually reach a county governing body, and the questions worth asking before the vote.
South Carolina already has a SCAIO primer on AI for local government. It is written for administrators, IT directors, and clerks — the people who select tools, configure them, and defend the network.
A county council member's job is different. You rarely choose the software. You approve the budget that buys it, the rezoning that hosts it, the incentive agreement that finances it, and the policies that govern it. You are accountable to residents for all four, usually in public, often with incomplete information, and frequently on a timeline set by someone else.
This primer covers the AI decisions that arrive on a county council agenda — and what to ask when they do.
Rezonings, siting, infrastructure commitments, and fee-in-lieu-of-tax agreements for data centers. The largest dollar figures a council will see, and the decisions hardest to reverse.
AI tools bought for county departments — drafting, permitting, call handling, assessment. You approve the spending and, in practice, set the policy floor for how it is used.
Public records, open-meeting materials, and what residents are entitled to know about how decisions affecting them were made.
SC Daily Gazette reported in March 2025 on "Moc-1," a proposed $2.8 billion computing center in Spartanburg County. According to that reporting, the facility would generate its own electricity on site, would serve the company's own internal operations rather than lease capacity, and is expected to employ 27 full-time staff.
The county's offer, as reported: a reduced 4% property tax rate for 40 years against the standard 6%, plus tax credits of roughly $1.5 million annually for the first 20 years.
None of that makes the deal wrong. On-site generation genuinely avoids the grid-impact questions other projects raise, and a large capital investment can carry a county's tax base for decades. But a 40-year commitment at 27 jobs is a different proposition from a 40-year commitment at 700 jobs, and it should be evaluated as the proposition it actually is.
Without an explicit price-signal mechanism, the cost of meeting that growth tends to flow through to ratepayers broadly — meaning the residential customer in a 1,200-square-foot home in Rock Hill helps subsidize the cost of grid expansion driven by a hyperscaler facility that uses more power in an hour than the home does in two years."
In April 2025, SC Daily Gazette reported that Santee Cooper voted to enact an experimental electricity rate structure for customers drawing 50 megawatts or more — a four-year pilot designed so that very large new loads contribute to the cost of new generating capacity rather than having those costs spread across existing residential and small-business customers.
The 50 MW threshold captures essentially every modern AI-focused data center campus, and several large EV-and-battery plants besides.
This is genuinely contested ground, and reasonable people land in different places. Some peer states subsidize large loads rather than surcharging them, betting that long-run economic benefit exceeds short-run ratepayer cost. Both arguments have merit. What matters for a council is knowing which bet your county is making, and on whose behalf.
Ask for it in writing. It is the single clearest way to compare an AI-infrastructure package against the industrial deals a county already knows how to evaluate.
On-site generation, a dedicated renewable facility, and ordinary grid draw are three very different answers with three very different bills attached.
Cooling water is the least-discussed input in most siting conversations and among the most locally consequential.
Every incentive agreement should answer this. If the draft in front of you does not, that is the amendment to ask for.
Most council members voting on a 40-year agreement will not be in office for the back half of it. Someone should still be able to describe what it obligates.
Non-disclosure is normal in economic development. It is still a decision, and residents are entitled to know its shape even when they cannot know its contents.
Whether or not a county has an AI policy, staff are using AI tools — drafting correspondence, summarizing long documents, preparing meeting materials. This is mostly good, and the productivity gains are real. It is also happening whether or not anyone has written down the rules.
Councils do not need to become technologists to govern this well. Two decisions cover most of it: what categories of county data may be put into an outside AI service, and which decisions affecting residents require a named human being accountable for the outcome.
For the procurement side of this, see SCAIO's primer on AI procurement for SC government.
The statewide member organization for SC's 46 counties. Model policies, training for newly elected officials, group purchasing, and peer-county comparison on incentive practice.
State-level convening on AI governance. A useful reference point when a county is drafting its own AI use and procurement guidance.
Threat intelligence, training, and incident response available to SC local governments — including the voice-cloning and impersonation risks aimed at elected officials specifically.
Independent, public-interest AI research for South Carolina. The flagship report, the policy tracker, and these primers are free to use and cite.
For the full infrastructure picture, read Chapter 5 of the flagship report and the Journal article "South Carolina's Data-Center Decade." For the operational side of county AI use, see the primer on AI for SC local government.
scaio.org · jimmy@scaio.org