A landscape survey is incomplete without naming the failure modes worth planning around and the leadership cases worth pursuing. This chapter does both — in the same constructive register as the rest of this report.
Every chapter to this point has tried to describe what is. This one looks at what could be — in both directions. Section one walks through six risks that South Carolina would benefit from planning around now rather than later. Section two walks through six opportunities where the state is plausibly positioned to lead nationally if the existing pieces are connected.
Neither list is a critique. The risks are not failures of the institutions that would carry the load; they are common challenges every state faces as AI moves from pilot to production. The opportunities are not promises; they are credible claims grounded in the assets South Carolina already has. The work of the next several years will be to mitigate the first list while pursuing the second.
Generative AI expands the cyberattack surface for South Carolina's public institutions and lowers the cost of social engineering at scale. The institutions most exposed are the ones with the largest information footprints relative to their security budgets: school districts and the State Department of Education, hospitals and rural health systems, municipal and county government, the state court system, and the Port of Charleston's digital infrastructure. Phishing, voice cloning, and synthetic-media impersonation are already being used against organizations of this profile elsewhere in the country.
South Carolina has structural advantages here. The state has an established cybersecurity community anchored in Charleston, Naval Information Warfare Center Atlantic, the SC Cyber Initiative, and a deep bench of practitioners at the Citadel, USC, Clemson, and in the private sector. SC ETV reaches every school district and could serve as a literacy channel. The constructive question is how to convert that existing capacity into an information-sharing and red-teaming function that smaller institutions can draw on without standing up dedicated programs of their own.
What would help: a coordinated AI-aware threat-intelligence channel for public institutions, baseline tabletop exercises that include synthetic-media scenarios, and shared procurement of AI-specific email and identity controls for school districts and rural hospitals — all of which could be assembled from capacity that already exists in the state.
Generative AI raises three distinct election-related risks: synthetic media depicting candidates or officials, AI-generated misinformation about voting procedures, and the use of AI to scale traditional influence operations. The risk is not unique to South Carolina — it is national — but every state's election apparatus has to adapt locally.
The State Election Commission already manages a complex, decentralized system across 46 counties. The agency's existing voter-education infrastructure is well positioned to incorporate synthetic-media literacy without standing up new programs. A constructive contribution from outside the agency would be to make plain-language, non-partisan guidance about how to recognize and report synthetic political content available through trusted channels — county libraries, civic associations, faith communities, the Press Association — well in advance of the next election cycle.
Two areas would benefit from clear, voluntary standards: candidate disclosure of AI-generated content in their own communications, and a defined contact path for campaigns and journalists to flag suspected synthetic content for review. Neither requires new legislation; both could be developed collaboratively between the State Election Commission, the Press Association, and the state's two major political parties.
The most underappreciated AI risk in South Carolina is that AI-augmented learning could widen the gap between well-resourced and under-resourced school districts rather than narrow it. Districts with the staff capacity to evaluate, deploy, and supervise AI tools will accelerate their students' use of them. Districts without that capacity may either ban tools their students will use anyway, or adopt them without the guardrails the well-resourced districts have built.
H.5253's parental-consent framework, discussed in Chapter 6, is one of the most thoughtful first-in-the-nation attempts at this problem. But consent alone does not equalize capacity. The under-recognized constraint is district-level expertise: someone in each district has to understand the tools well enough to write the policy, train the teachers, and answer parents' questions.
What would help: a state-level AI-in-education resource hub that any district can draw on without contracting; a shared model policy and training curriculum that can be adapted rather than written from scratch; and an explicit pathway for technical colleges and universities to provide capacity-building support to under-resourced districts. The South Carolina Department of Education and the SC ETV K-12 platform are well positioned to anchor that work.
S.443 (Health Claims & AI) addresses one slice of healthcare AI — insurer prior authorization — and does so with care. The larger set of clinical-AI deployment questions remains open: AI-driven diagnostic triage in rural hospitals, deployment of ambient documentation tools (such as DAX Copilot, already in use at MUSC), AI-assisted radiology and pathology, patient-data governance when models are trained on EHR data, and equitable access to AI-enhanced care across the state's health systems.
The risk is two-tiered. At the high end of the system, MUSC, Prisma Health, and the larger systems have the resources and governance structures to deploy AI carefully. At the rural-hospital end, deployment decisions may be made by vendors and consultants rather than by clinicians and patients. The gap between those two tiers is where a thoughtful statewide approach could be most valuable.
What would help: a voluntary statewide clinical-AI deployment standard that incorporates evidence requirements, bias auditing, and informed-consent practices — drawing from MUSC's existing AI governance, the work of the Clemson–MUSC AI Hub, and rural-hospital perspectives. South Carolina would not be the first state to attempt this, but the existing institutional alignment in the state could make it among the most credible.
Chapter 4 walked through the two-track exposure pattern in South Carolina's workforce: cognitive office and customer-service occupations that face augmentation pressure today, and the broader bench of operational, manufacturing, and skilled-trade occupations where AI's effects arrive more slowly but more durably. The risk is that the speed of change in the augmentation-track occupations outpaces the speed of South Carolina's re-skilling infrastructure.
The state's Technical College System is the most underused asset in this picture. Sixteen technical colleges, every county within commuting distance of at least one, and an existing infrastructure for short, credential-bearing programs — but limited AI-specific pathways at present. Re-skilling speed is fundamentally a question of how fast a worker can move from displaced to differently employed; technical colleges have the structural ability to compress that timeline if the programs exist.
What would help: short, stackable AI-adjacent credentials targeted at the operational roles that AI deployments actually create (data labeling and quality assurance, MLOps support, applied prompting and workflow integration, AI-aware customer service, model-output review). The technical college system, the Department of Employment and Workforce, and the SC ReadySC employer-customized training program could anchor this work together.
A small number of large institutions account for most of South Carolina's identifiable AI capacity: USC's AI Institute, the Clemson School of Computing, the Clemson–MUSC AI Hub, the MUSC Center for AI, SCRA's AI Leadership Hub, and a handful of large enterprise deployments at BMW, Boeing, and the major health systems. That concentration is partly a feature — it produces the kind of institutional alignment that other states would envy — and partly a vulnerability.
The vulnerability shows up two ways. First, smaller communities — the rural Pee Dee, the Lowcountry outside Charleston, the Upstate outside Greenville — are further from the institutional capacity that drives access to grants, talent, and tools. Second, the ecosystem's resilience depends on a small number of leadership decisions: a change in priorities at any one of the anchor institutions would meaningfully change the state's AI posture.
What would help: lightweight participation pathways for smaller institutions — small-town libraries, regional community colleges, county economic development offices, faith-based organizations doing community technology work — so that the ecosystem's center of gravity is distributed even if the institutional anchors remain concentrated. The Charleston Digital Corridor's regional model offers one template; ADAPT in SC's NSF-funded EPSCoR network offers another.
South Carolina has assembled an unusual concentration of advanced manufacturing in a state its size: BMW's largest manufacturing plant by volume globally (Spartanburg), Boeing's 787 final assembly line (North Charleston), Volvo Cars' US plant (Berkeley County), Mercedes-Benz Vans (Charleston), Michelin's North American headquarters and multiple plants, the Bridgestone tire plant, the Continental Tire plant, and a deep tier-one supplier ecosystem. Layered on top of that is Clemson's International Center for Automotive Research (CU-ICAR) in Greenville and USC's McNair Aerospace Center. The applied-AI capability built on top of that base is what creates the leadership claim.
The most visible signal is BMW's pilot deployment of Figure AI's humanoid robots at the Spartanburg plant — among the earliest commercial humanoid-robotics deployments anywhere. That is not yet a productivity result, but it is a credible signal that SC's manufacturing base is being treated as a serious applied-AI laboratory by national-tier vendors. The opportunity is to extend this beyond one pilot at one plant into a state-level applied industrial-AI consortium that links Clemson and USC research capacity with the manufacturer base.
The constructive frame: South Carolina does not need to invent industrial AI. It needs to convene it. A standing industry-academic forum with a small, focused agenda — shared workforce pipelines, joint research projects on specific applied problems, a public showcase of what the state is already doing — could turn a strong assembly of capabilities into a recognizable national position.
Charleston's defense, cybersecurity, and intelligence-community presence creates an opportunity few other states can match. Naval Information Warfare Center Atlantic, SPAWAR's successor, anchors a community of defense contractors, cyber firms, and cleared engineering talent that runs from North Charleston up the I-26 corridor. SCRA's existing AI Leadership Hub already brings together public-sector, academic, and industrial actors with a defense and cyber slant.
The opportunity is at the intersection of AI and cybersecurity — both AI-enabled defensive tooling and the secure development and deployment of AI systems for sensitive government use. Few states combine the cleared workforce, the defense-industrial base, the academic capacity, and a coherent public-sector champion in one place. The constructive observation: the pieces exist; what would help is a recognizable convener with a defined remit to surface SC's work in this area to the federal customer base.
The Port of Charleston is one of the country's largest and fastest-growing container ports, with the Inland Port at Greer extending its reach deep into the Upstate manufacturing corridor. The SC Ports Authority has been digitizing operations for years — appointment systems, intelligent gate operations, terminal management — and is well positioned to be among the AI-forward ports in the country over the next decade.
Port AI is one of the application areas where productivity gains are most concrete and least speculative. Predictive maintenance on cranes and yard equipment, intelligent berth scheduling, container-stack optimization, and real-time logistics coordination with inland-port operations are well-defined problems with established solutions. The opportunity is to position the SC Ports Authority's continued digitization as a national reference case — a public-sector entity demonstrating disciplined, results-oriented AI deployment.
Clemson's Precision Agriculture program and Clemson Cooperative Extension service give South Carolina a stronger applied-agriculture research base than its land area suggests. Layered on top of the state's recent broadband investment — including significant federal infrastructure funding directed at rural connectivity — there is a credible path to making SC farms among the most data-rich in the country.
Precision agriculture is one of the rare areas where AI's productivity benefits flow most directly to smaller operators: variable-rate fertilization, automated irrigation scheduling, pest and disease detection, livestock-health monitoring. Where larger commercial operations have led adoption nationally, SC's mix of mid-sized family farms and the existing Cooperative Extension delivery network provides a natural channel for broad participation rather than a winner-takes-all dynamic.
The constructive observation: agricultural AI is one of the areas where state-level convening matters more than state-level funding. The Department of Agriculture, Clemson Extension, the SC Farm Bureau, and the State Conservation Districts already work together on adjacent issues. Adding AI-specific guidance, demonstration plots, and farmer-to-farmer learning networks would extend infrastructure that already works.
MUSC's Center for AI, the joint Clemson–MUSC AI Hub, and Hollings Cancer Center (an NCI-designated comprehensive cancer center) collectively make MUSC one of the most active academic medical centers in the country for applied clinical AI. MUSC's appointment of a Chief AI Officer, the deployment of DAX Copilot for ambient clinical documentation, and the Pap-smear AI work in oncology have all drawn national attention.
The opportunity is to translate institutional leadership into a recognizable statewide model — a deployment standard, governance template, and evaluation framework that smaller SC health systems and rural hospitals can adopt without rebuilding it themselves. Few states have an academic medical center with this much applied capacity paired with a clearly defined rural-health context to translate it into.
The combination of a published AI Strategy (Department of Administration), the AI Center of Excellence, the SCRA AI Leadership Hub, and active university institutes gives South Carolina an unusually structured starting point for AI in public service. Most states do not have any of these; many that have one do not have all four.
The opportunity is to compound those assets into a state-government AI literacy standard that other states would borrow. Three audiences need different curricula calibrated to the same vocabulary: legislators (what AI is, what it can do, what the policy levers are), agency staff (what AI is, how to use it responsibly in their workflows, what the procurement and security questions are), and the public-facing communications channels of state government (how to talk about AI accurately to constituents). All three could be assembled from work the state has already done or commissioned.
Chapter 8 takes the six opportunities and the six risks and converts them into a constructive set of recommendations for the institutions doing the work. Chapter 9 names the sources behind the analysis. None of these recommendations are prescriptions; they are observations about what would help, written from a position of championing the work already underway.