Developer Synthetic Data Adoption & Ethics Survey

Welcome! Thank you for participating in this survey about synthetic data practices. This survey takes approximately 11 minutes to complete. Your participation is entirely voluntary, and you may stop at any time. There are no right or wrong answers — we are interested in your honest opinions and experiences. All responses are confidential, anonymized, and reported only in aggregate. Results will be used for internal research to better understand developer needs around synthetic data.

In the last 12 months, have you worked with synthetic data?

• Yes, regularly (monthly or more)
• Yes, occasionally
• No, but I am familiar with the concept
• No, and I am not familiar with it

Which tools or approaches have you used to generate synthetic data? (Select all that apply)

• In-house generation scripts
• Open-source libraries (e.g., SDV, SynthCity, ydata-synthetic)
• Vendor platform (e.g., Gretel, Mostly AI, Tonic)
• Data augmentation utilities not aimed at privacy
• Haven't used tools directly (consumed output from others)
• Other (please specify)

How much demonstrated fidelity and utility do you require before using synthetic data in production?

How likely are you to increase your use of synthetic data in the next 6 months?

Based on your responses in this survey, please share any additional thoughts about your limits, ideal use cases, or expectations for synthetic data.

Which best describes your primary role?

• Software engineer
• ML/AI engineer
• Data scientist
• Data/ML platform engineer
• Security/privacy engineer
• Product or engineering manager
• Researcher/academic
• Other (please specify)

Thank you for participating! Your input helps us understand practical needs and considerations around synthetic data. If you have any questions about this research, please contact [research team email].

<p>For this survey, <strong>synthetic data</strong> refers to artificially generated data (e.g., via simulations or generative models) intended to mimic real data's statistical properties while protecting sensitive information or filling gaps.</p>

Which use cases for synthetic data are most relevant to you or your team? (Select all that apply)

• Prototyping or training ML models
• Class imbalance augmentation
• Privacy-preserving sharing or compliance
• Testing and QA (e.g., edge cases, rare events)
• Synthetic logs or telemetry for load testing
• Analytics demos or sandboxing
• Education or training
• Other (please specify)

Approximately what percentage of data in your projects over the last 12 months was synthetic?

• 0% (none)
• 1–10%
• 11–25%
• 26–50%
• 51–75%
• 76–100%
• Not sure

<p>How appropriate is synthetic data for <strong>model training and development</strong> in your context?</p>

Which factors most limit your use of synthetic data today? (Select all that apply)

• Hard to evaluate quality or metrics
• Limited domain coverage
• Tooling or integration gaps
• Compute or cost constraints
• Stakeholder skepticism or buy-in
• Policy or legal uncertainty
• No clear need
• Other (please specify)

We'd like to explore a few of your responses in more depth. An AI moderator will ask you up to 2 brief follow-up questions based on what you've shared so far.

How many years of professional experience do you have in software or data roles?

• 0–1 years
• 2–4 years
• 5–9 years
• 10–14 years
• 15+ years

<p>How appropriate is synthetic data for <strong>production decision-making</strong> in your context?</p>

Rank the following improvements by how much they would accelerate synthetic data adoption in your organization, from most to least impactful.

1. Better quality and validation metrics
2. Broader domain and data type coverage
3. Easier integration with existing pipelines
4. Lower cost or compute requirements
5. Clear policy and legal guidance or templates
6. Independent benchmarks and case studies
7. Training and best-practice playbooks

What is your primary domain or industry?

• Technology
• Finance/FinTech
• Healthcare/Life sciences
• Retail/Consumer
• Telecom/Media
• Manufacturing/Industrial
• Government/Public sector
• Education
• Other

<p>How appropriate is synthetic data for <strong>testing and QA</strong> in your context?</p>

What is your organization's approximate size (global headcount)?

• 1–9
• 10–49
• 50–249
• 250–999
• 1,000–4,999
• 5,000–19,999
• 20,000+

<p>How appropriate is synthetic data for <strong>external reporting or compliance submissions</strong> in your context?</p>

In which region do you primarily work?

• North America
• Latin America
• Europe
• Middle East
• Africa
• South Asia
• East Asia
• Southeast Asia
• Oceania

Rank your top concerns about synthetic data from most to least concerning.

1. Privacy leakage or re-identification
2. Bias amplification or fairness issues
3. Poor realism or utility
4. Regulatory or compliance risk
5. Lack of transparency or traceability
6. Leakage of secrets or intellectual property

If compliance or privacy is a priority in your work, briefly describe the data types or regulations you must satisfy (e.g., HIPAA, GDPR, PCI-DSS). If not applicable, you may skip this question.