AI in Healthcare (Part 2) - Why Scaling Is Still So Hard
Based on the World Economic Forum & BCG’s January 2025 White Paper: “The Future of AI-Enabled Health”.
In Part 1, we talked about the promise of AI in healthcare.
Now let’s talk about what’s holding it back.
While pilots and prototypes are everywhere, very few AI solutions have actually scaled across health systems.
The World Economic Forum and Boston Consulting Group’s January 2025 white paper lays it out clearly:
There are two categories of barriers—some we can fix, and some that are much harder to move.
Two Categories of Barriers
Experts break the blockers into:
Structural constraints – fixed and foundational
Challenges that can be solved – with the right collaboration
Let’s look at both.
1. Structural Constraints
These are deeply embedded issues in how health systems operate. They’re difficult to change and often require systemic reform over time.
A. Political Conditions
Election cycles create pressure to deliver short-term results within 2 to 3 years.
But AI in healthcare needs long-term commitment.
Scaling innovation is often limited by national systems that aren’t designed for this kind of sustained effort.
B. Need for Scale
To show real impact, AI in health must be validated at scale.
Small pilots don’t cut it.
Public health outcomes need large-scale, long-term data to prove that AI makes a difference.
C. Resource Limits
Budget pressures are a global issue.
Organisation for Economic Co-operation and Development (OECD) countries already spend around 11% of GDP on healthcare.
In many cases, health spending is growing faster than the overall economy — leaving little room for new investments like AI.
D. Resistance to Change
There’s an inherent resistance to change within the healthcare sector.
This includes both inertia and conservatism — especially in clinical and administrative settings where trust in new systems takes time to build.
E. Legacy Systems and Processes
Most health systems still rely on outdated IT, regulations, and incentive structures.
These legacy elements slow down the adoption of AI, even when the technology itself is ready.
2. Challenges That Can Be Solved
With strong public–private partnerships, these barriers can be overcome.
A. AI Feels Too Complex for Leaders
AI in health is attracting attention, but many leaders struggle to prioritise it. Why?
There’s no clear, shared strategy that links AI to public health goals
Policymakers don’t always understand where AI fits
Coordination is messy and financing is unclear
This complexity stops AI from being placed high on the agenda.
B. Tech Decisions Aren’t Aligned with the Bigger Picture
Often, AI decisions are left entirely to tech teams.
But without alignment with broader healthcare strategies, it leads to:
Unequal access to AI tools due to inadequate DPIs (digital public infrastructure)
Poor digital literacy among public leaders and decision-makers
This disconnect holds everything back.
C. Trust Issues and Patchy Governance
As AI becomes more common, people want to know:
Is it fair? Safe? Regulated?
But today’s governance is:
Fragmented
Outdated
Hard for the public to trust
Without clear, transparent systems, trust becomes a barrier.
Putting It All Together
These aren’t just tech problems.
They’re also about:
How decisions are made
How systems are structured
And how much people trust what’s being built
To make real progress, we need to:
Bring leadership and vision together
Match strategy with on-the-ground execution
Build smarter, more open governance
And invest in digital literacy - not just technology
Coming Up in Part 3
In the next issue, I’ll break down:
Four visions for AI-enabled health
Source
This substack article is based on: “The Future of AI-Enabled Health” – January 2025, by the World Economic Forum and Boston Consulting Group.
Link to the report - Click Here
Until next time, Keep Smiling,
Arjun Shrivastava,
Founder – Curaa.in ,