Building HIPAA-Compliant AI Automations: AWS vs. Google Cloud

Start here — the decision you're actually making

You're choosing where to build AI automations that touch patient data. Read this framing first.

An "AI automation" here means software your team builds that calls an AI model through an API (a programmatic request — e.g., "summarize this note," "extract the diagnoses," "answer this patient question") and chains those calls into a workflow or agent (software that reasons over several steps and uses tools on its own). If that workflow ever touches PHI (Protected Health Information — patient data tied to a person), it falls under HIPAA, and the platform underneath has to support that. Both AWS and Google Cloud do.

Three ground rules shape everything below:

• No cloud — and no AI model — is "HIPAA certified." There's no such stamp. AWS says it plainly: "There is no HIPAA certification for a cloud service provider."¹ Google says the same — "there is no certification recognized by the US HHS for HIPAA compliance."⁵ Be skeptical of any AI vendor claiming to simply be "HIPAA compliant" out of the box.
• It's a shared responsibility. The cloud secures the infrastructure and the model service; your team is responsible for configuring it correctly and only sending PHI to approved services in an approved way.^{1 5} This is where most AI-automation compliance mistakes happen — see the #1 risk.
• The BAA is the gate. Before any PHI flows through your automation, your organization signs a BAA (Business Associate Agreement), the contract in which the vendor commits to protect that data. Both clouds make it self-service.^{1 5}

The 30-second answer

If you read one section, read this one.

Both AWS and Google Cloud are fully capable platforms for building HIPAA-compliant AI automations. On both, your team can sign a BAA in minutes,^{1 5} call top-tier AI models — including Anthropic's Claude — by API under that BAA,^{2 5} build multi-step agents with a managed framework, and do it with no special "HIPAA tax."⁵ For most teams, the right platform is the one your engineers already build on, not the one that wins a feature checklist.

The differences that actually matter for AI automations are narrow, and each platform has a distinct strength:

• AWS gives you a precise, enforceable per-model data-retention control (set it to "keep nothing"),³ and a deeper bench of ready-made clinical AI, such as ambient note-taking and clinical-text extraction.²
• Google Cloud gives you a simpler scope to reason about (its BAA spans the whole infrastructure),⁵ a standing written commitment that your data won't be used to train its models,⁶ and an explicit written promise of no HIPAA price premium.⁵

If you're already building on one of them, the strong default is to stay — you should not switch clouds for compliance reasons alone.

Side-by-side comparison

Six dimensions that matter when building AI automations on patient data. Each cell gives the finding plus what it means for you, cited to its source. Genuine ties are labeled honestly.

Tally: one slight edge to AWS (prebuilt clinical AI), five genuine ties. Neither platform is a "winner" for building HIPAA AI automations.

1 · The BAA gate

The first thing your organization does on either cloud, before any PHI moves through an automation.

The BAA is the vendor's legal commitment to protect the patient data you send it. On both clouds it's self-service — no procurement cycle, no lawyer-to-lawyer negotiation for the standard agreement. On AWS you accept it in a console tool called AWS Artifact, and a single acceptance can apply to every account your organization manages.¹ On Google Cloud you accept it in the console.⁵

On both clouds, signing the BAA does not bless the entire platform for PHI. AWS gates PHI to a published HIPAA-eligible services list;² Google's BAA covers its infrastructure and a published list of covered products. In practice both work the same way — check the list, keep PHI on it.

What it means for you: a tie. Either way, your team can have the agreement in place the same day, at no cost.

2 · Calling AI models by API — the core of any AI automation

Can you make API calls to strong models, under HIPAA? On both, yes — including Claude.

• AWS — Amazon Bedrock. A single API/SDK to call many foundation models, including Anthropic's Claude, Meta's Llama, and Amazon's own models. Bedrock is HIPAA-eligible.²
• Google Cloud — Gemini Enterprise Agent Platform. API access to Google's Gemini models plus a model garden that also includes Claude.⁵

What it means for you: a tie on raw capability. If your team specifically wants Claude under a BAA, you can have it on either platform. Decide on the rest of your stack and on the data-handling details in section 4.

3 · Building and running agents

Beyond one API call: frameworks to chain steps, use tools, and run agents in production.

• AWS — Bedrock AgentCore, a HIPAA-eligible framework for building and deploying agent workflows, paired with Bedrock Guardrails, which can automatically detect and mask sensitive identifiers in prompts and responses.²
• Google Cloud — Agent Development Kit (ADK) for building multi-agent systems, with a managed Agent Engine runtime to deploy and scale them. The agent runtime supports HIPAA, customer-managed encryption keys, and network isolation controls.⁵

What it means for you: a tie. Both are credible, production-grade agent platforms.

4 · The #1 PHI risk — what happens to your prompts

This is where AI-automation compliance is won or lost. "The platform is HIPAA-eligible" does not mean every model is safe for PHI by default.

When your automation sends a prompt to a model, you must know: does anyone store or see that prompt afterward? A prompt can contain PHI, so retention and provider-sharing are the core HIPAA questions.

AWS — an explicit retention switch ³

• none — zero data retention. No request or response data is stored by AWS or shared with the model provider. This is the setting to use for PHI.
• provider_data_share — AWS retains your prompts/outputs and shares them with the model maker. Required by certain models. Avoid for PHI.

If you've set none and your code calls a model that requires sharing, that model shows as unavailable and the request is rejected rather than leaking data. You can enforce none across the whole organization with a Service Control Policy. The trade-off: it's a switch your team has to set correctly — get it wrong and the protection isn't there.³

Google Cloud — a standing commitment plus controls ⁶

Under the Google Cloud AI/ML Privacy Commitment, "customer data … is not used to train foundation models."⁶ Important nuance: that is a training commitment — confirm the retention terms for the specific service you call. On top of the commitment you add customer-managed encryption keys and network isolation controls.^{6 5}

What it means for you: a tie on outcome, reached two different ways. Both can be run compliantly; both punish inattention.

5 · Prebuilt healthcare building blocks

How much clinical capability you get without building it yourself.

• AWS has the deeper set of ready-made clinical AI, all HIPAA-eligible: HealthScribe (turns a visit conversation into a draft note), Comprehend Medical (extracts diagnoses, medications, and procedures from free-text notes), and HealthLake (a clinical data store using FHIR).²
• Google Cloud offers the Cloud Healthcare API and Healthcare Data Engine — strong, standards-based services for ingesting and managing healthcare data (FHIR, DICOM, HL7v2) — but fewer turnkey "clinical AI" features.⁵

What it means for you: a slight edge to AWS if you want clinical AI capabilities working out of the box.

6 · Cost & support

Is there a "HIPAA tax," and who picks up when your automation breaks?

Cost: neither charges a premium for HIPAA. Google states it outright — "we can offer HIPAA regulated customers the same products at the same pricing … Other public clouds charge more money for their HIPAA cloud, we do not."⁵ AWS charges no fee for the BAA and no surcharge for regulated workloads.¹

Support: comparable ladders. AWS paid tiers are 24/7, with urgent-issue response targets from under 30 minutes to under 5; a named Technical Account Manager comes with Enterprise.⁴ Google Cloud offers Standard, Enhanced, and Premium, with a named Technical Account Manager at Premium.⁷

What it means for you: a tie. Google puts the price promise in writing.

Price & confirm for your specific build

Compliance is broadly a tie. These build-time variables often decide the real cost and feasibility.

• Model token pricing. Inference is usually the dominant cost at scale. Compare the specific models you'll use — prices are not guaranteed to match across platforms.
• Model & region availability under the BAA. Confirm the model versions you want are available and BAA-eligible in your required region.
• Where inference runs (data residency). When a PHI-containing prompt is sent, which region processes it? Verify you can pin it to your required geography.
• Rate limits & latency. Real-time clinical workflows can hit throttling. Check each platform's rate limits and typical latency for your call volume.
• Audit logging. Both clouds log activity, and both can log model calls — but prompt/response logs can contain PHI, so those logs must be encrypted and access-controlled.
• Fine-tuning / bring-your-own-model. If you plan to fine-tune on clinical data, confirm that path is HIPAA-eligible and understand where the training data and resulting model live.

Which is the better fit if…

No single winner — but clearer answers once you know your situation.

Before you build — five things to confirm

This guide is a starting map, not a compliance sign-off. Verify these with your team or vendor.

• Check each model's data handling first. On AWS, set retention to none and confirm the model supports it;³ on Google, confirm the service is BAA-covered, check its retention terms, and apply customer-managed keys + network controls.^{5 6}
• "Eligible/covered" is not "compliant." A service being on the approved list only means you're allowed to use it with PHI once your BAA is signed and it's configured correctly.
• The lists and brands change. Both vendors update their approved-service lists regularly — re-check the live pages when you build.
• Price out the build-time variables. See Price & confirm for what actually decides cost and feasibility.
• This is not legal advice. Final HIPAA sign-off belongs with your legal, privacy, or compliance owner.

Sources

Every numbered citation above links here. All pages were read directly on 2026-06-12.