Artificial intelligence has rapidly transformed health care in many ways, including improving diagnoses, automating clinical documentation, and streamlining administrative tasks. For employers in particular, AI has potential to simplify benefits administration and lead to better data insights. Alongside this promise, employers must ask tough questions to ensure that AI solutions advance value, as opposed to inflating costs and potentially creating indeterminate levels of risk.

AI’s Current Use

Across the workplace, employers have started to leverage AI as they reimagine the employee benefits experience. Currently, employers have leveraged AI applications that include:

• Human-like avatars that provide 24/7 benefits guidance and allow employees to make informed decisions at any time;
• Personalized benefits recommendations, tailored to an employee’s family, health status and usage patterns;
• Mental health and well-being tools that offer scalable, ongoing support beyond traditional working hours;
• Automated document processing that boosts compliance, benefits administration and claims review; and
• Global platforms that centralize benefits information and produce insights from varied data sources.

As with any emerging technology, the potential value of any AI application must be weighed against the costs, complexity and governance required to responsibly deploy it.

What Employers Should Know About AI and Their Vendor Partners

Industry partners have also broadly embraced artificial intelligence, though some have labeled products as “AI-powered,” while misrepresenting or exaggerating the actual use of AI, a practice known as “AI washing.” Employers must demand transparency from their partners concerning AI, including whether it improves outcomes while truly lowering their costs.

As employers hold vendor partners accountable for artificial intelligence, key questions include:

• Do the tools result in reductions, or worse, increases in fees or claims costs?
• Does the vendor demonstrate both technical expertise in AI and a deep understanding of the health benefits environment?
• What specific AI models are being used, and how is data privacy managed and secured?
• How well does the solution integrate with existing internal and third-party HR, benefits, well-being and/or payroll systems?
• Can the vendor clearly articulate how AI advances the solution and provide real-world performance data?
• How is success of the AI tool measured and against what baseline?
• How does this AI application impact the member experience?
• What testing does the tool undergo to ensure it meets proper safety standards for the purpose which it’s being used?

When vetting new solutions, employers should look beyond marketing claims and seek evidence of measurable impact and responsible governance.

The Use of AI in Clinical Innovation

Clinical applications of artificial intelligence are proliferating quickly. Understanding their influence on health care delivery and payment is essential for making informed benefits decisions, especially surrounding cost transparency, equitable outcomes and data security.

AI already powers diverse clinical applications that include taking notes during visits, known as “scribing,” and leveraging virtual assistants to support care coordination. AI is also helping to diagnose conditions, and predictive modeling tools are being used to interpret radiology images and assess disease progression and potential treatment success. Some clinical practices are using AI for billing assistance, which has the potential to increase claim costs.

Employers should press their partners on their clinical AI strategies to evaluate both the potential for better health outcomes and to ensure safeguards are in place.

Guardrails for Responsible AI Adoption

Employers will want to adopt AI in a deliberate and phased manner to minimize risk, ensure effective integration and protect employee trust. That includes streamlining and digitizing benefits content so it is searchable and ready for machine learning, piloting tools that support navigation and communication with clear human oversight, exploring AI-enabled claims monitoring and tracking outcomes so that strategies can be refined over time.

At the same time, employers must be clear about what they are implementing and where lines should be drawn. Any employer-sponsored AI solution should be governed by robust safeguards for HIPAA, General Data Protection Regulation (GDPR) and other applicable frameworks, with explicit limits on use cases so that tools answer questions like deductibles or networks but do not provide clinical guidance such as advice on medication side effects. Bias in algorithms and variability in responses also require careful risk stratification, clear accountability and human review, so AI augments benefits delivery rather than making de facto medical decisions.

While early changes may be somewhat incremental, the long-term impact of AI on employer-sponsored health care will certainly be profound. By thoughtfully embracing innovation and keeping employee trust at the forefront, employers can help shape a more accessible, efficient and optimized future for health care.