What is difference between AI safety and AI security?

Safety = protect people from the model (accidental harm: bias, hallucinations, leaking PII, over-reliance) — guardrails. Security = protect the model from attackers (adversarial: prompt injection, breaches, exfiltration). Intent is the key distinction.

No. Security teams need a security cert — look at TAISE ($795) or CASO (~$2,199). This is AI Safety for people who use AI on the job, not a CISO track.

Is this for researchers?

No. This is not a research seminar, fellowship on-ramp, or research career. It is a workplace credential for the job you already have if you use AI at work.

Is this a research course or a security cert?

Neither. CASF™ Foundations is AI Safety and Risk Management for people who use AI on the job. Security certs (TAISE) protect the model from attackers. Research literacy programs are selective cohorts with a completion certificate — not an identity-verified exam and not a public verify URL. Write the policy. Sit the identity-verified exam.

Is this for everyone with ChatGPT?

Only if AI is part of your job. PMs, engineers, lawyers, HR, and founders who use AI on the job — not a casual-use course for everyone.

Does this satisfy EU AI Act Article 4 AI literacy?

Article 4 requires providers and deployers to take measures to ensure a sufficient level of AI literacy among staff, tailored to their role, the context, and the risk of the AI systems they operate. Get AI Safety Certified Foundations is plain-English literacy and deployer-practice training covering oversight, transparency, and record-keeping. It is not an official EU, government, or accredited certificate, and it does not by itself meet every high-risk obligation (those expand in August 2026). Use it as role-tailored literacy, then map remaining duties with counsel.

Yes. Live classroom videos, no login, 90 minutes of Parts 1–3, plus the policy template. Write and quiz stay in your classroom. M1–M5 unlock at $199 after the 80% quiz. Passing the quiz does not by itself unlock paid modules.

Live now — AI Safety ≠ AI Security · for people who use AI on the job

AI Safety is not AI Security. Most teams only learn AI Security.

For people who use AI on the job. AI Security protects the model from attackers. AI Safety protects people, clients, and your company from what the model does.

Foundations $199vs $795 TAISE | vs $2,199 CASO

Start free M0 Create account

Watch Part 1 here — no login. Create an account to write the policy and take the quiz. M1–M5 unlock at $199.

12h self-pacedVerifiable badge at https://www.getaisafetycertified.com/verify/{id} No prerequisitesCASF™ Foundations Level

CASF™ — AI Safety Certified — Foundations Level

SAMPLE • NOT A CREDENTIAL

This certifies that Sample Holder has completed 12 hours of training and passed an identity-verified exam in AI Safety & Risk Management — Foundations Level

Issued

May 1, 2026

Valid Until

May 1, 2028

This certificate evidences completion of CASF™ — AI Safety Certified — Foundations Level and passing of an identity-verified assessment administered by Gabby Software Engineering LLC. The holder name is the legal name from ID check. Professional certificate, not ANAB-accredited personnel certification under ISO/IEC 17024, nor US Gov-recognized. Designation: CASF™ (Foundations). Verify: https://www.getaisafetycertified.com/verify/demo — Issued: May 1, 2026 — Valid Until: May 1, 2028

Sample layout — not a live badge

See sample certificate · Verify a real IDSAMPLE · https://www.getaisafetycertified.com/verify/demo

Free preview · M0 Part 1

5 min

Play Module 0 · Part 1

Safety — Why safety is not security

Watch the 5-minute preview here, then finish M0 free.

For people who use AI on the job — not CISOs

Three doors in. Same Safety vs Security core.

Door A

For PMs & Engineers

Pain: The bot hallucinated. Legal blamed me.

Seven failure modes plus a RAG/agent checklist you can take into the next ship review.

Door B

For Legal / Compliance

Pain: EU AI Act deployer obligations are unclear.

EU AI Act and NIST AI RMF in plain English for deployers — not CISO jargon.

Door C

For HR / Founders

Pain: Staff pasting client data into ChatGPT.

A 1-page AI Safety Policy HR will sign, plus a what-never-goes-in-a-prompt checklist.

Blog

AI safety, in plain English

All posts

Safety vs Security

AI Safety vs AI Security: Intent Is the Key Distinction (2026 Guide)

AI Safety vs AI Security for people who use AI on the job: accidental harm vs attackers. Intent is the split.

EU AI Act

EU AI Act Deployer vs Provider: AI Safety Training for Teams

Most people who use AI on the job are EU AI Act deployers, not providers. AI safety training starts from that split.

AI Policy

What Never Goes in a Prompt: Write the Policy (Samsung)

AI safety training for people who use AI on the job: source code, customer files, and secrets never go in a prompt. Write the policy.

Security for CISOs

TAISE $795 · 20–30h

Security governance: protect the model from attackers. Built for CISOs and security leaders. CASO is ~$2,199 on the same track.

Safety for builders

CASF™ $199 · 12h

Plain-English Safety: protect people from the model. Identity-verified exam, verifiable badge + LinkedIn + QR. For PMs, lawyers, HR, compliance, engineers, and founders.

Full comparison

Workplace AI Safety

Not a research course. Not a security cert.

CASF™ Foundations is AI Safety and Risk Management for people who use AI on the job. Security certs protect the model from attackers. Research and fellowship seminars are a different path: apply, cohort hours, a completion certificate, usually a career plan. We start with free Module 0, then $199 — write a workplace AI Safety Policy, sit an identity-verified exam, and publish a public verify URL.

Compare tracks

By Get AI Safety Certified Team, AI Safety Certified · September 2026

Key Takeaways

Safety = protect people from the model (accidental harm: bias, hallucinations, leaking PII, over-reliance) — guardrails. Security = protect the model from attackers (adversarial: prompt injection, breaches, exfiltration). Intent is the key distinction.
Get AI Safety Certified Foundations is $199 for every candidate. TAISE is a $795 security track for CISOs.
Free M0 is 90 minutes of public video plus a policy template. Write and quiz stay in the classroom.

What is AI safety certification?

An AI safety certification proves a builder can protect people, clients, and the company from what the model does — then pass an identity-verified exam. Get AI Safety Certified Foundations is that path in 12 hours and $199, with a verifiable QR + LinkedIn badge.

What is the difference between safety and security?

Safety mitigates accidental harm (bias, hallucinations, leaking PII, over-reliance) with guardrails. Security mitigates adversarial threats (prompt injection, breaches, exfiltration) by protecting the model from attackers. Intent is the distinction.

Do I need EU AI Act training as a deployer?

If your company uses AI at work, you are probably a deployer, not a provider. Article 4 asks for role-tailored AI literacy. This path is that literacy training — not an official EU certificate.

AI safety certification FAQ

What is difference between AI safety and AI security?: Safety = protect people from the model (accidental harm: bias, hallucinations, leaking PII, over-reliance) — guardrails. Security = protect the model from attackers (adversarial: prompt injection, breaches, exfiltration). Intent is the key distinction.
Is this for CISOs?: No. Security teams need a security cert — look at TAISE ($795) or CASO (~$2,199). This is AI Safety for people who use AI on the job, not a CISO track.
Is this for researchers?: No. This is not a research seminar, fellowship on-ramp, or research career. It is a workplace credential for the job you already have if you use AI at work.
Is this a research course or a security cert?: Neither. CASF™ Foundations is AI Safety and Risk Management for people who use AI on the job. Security certs (TAISE) protect the model from attackers. Research literacy programs are selective cohorts with a completion certificate — not an identity-verified exam and not a public verify URL. Write the policy. Sit the identity-verified exam.
Is this for everyone with ChatGPT?: Only if AI is part of your job. PMs, engineers, lawyers, HR, and founders who use AI on the job — not a casual-use course for everyone.
Does this satisfy EU AI Act Article 4 AI literacy?: Article 4 requires providers and deployers to take measures to ensure a sufficient level of AI literacy among staff, tailored to their role, the context, and the risk of the AI systems they operate. Get AI Safety Certified Foundations is plain-English literacy and deployer-practice training covering oversight, transparency, and record-keeping. It is not an official EU, government, or accredited certificate, and it does not by itself meet every high-risk obligation (those expand in August 2026). Use it as role-tailored literacy, then map remaining duties with counsel.
Is M0 really free?: Yes. Live classroom videos, no login, 90 minutes of Parts 1–3, plus the policy template. Write and quiz stay in your classroom. M1–M5 unlock at $199 after the 80% quiz. Passing the quiz does not by itself unlock paid modules.

Start free M0 EU AI Act for deployers

Certification Path — One Brand, Three Levels

Start at Foundations — no prerequisites. Stack to Professional and Expert as you grow. One verifiable brand: CASF™.

CASF™ — AI Safety Certified

Foundations Level — OPEN

12h self-paced • 6 modules • Quizzes
60Q identity-verified exam • 70% pass • 2 attempts
Capstone: Write org AI Safety Policy
Verifiable badge at https://www.getaisafetycertified.com/verify/{id}

$199

Unlock Foundations — $199

No prerequisites. Free M0, then $199 for every candidate.

🔒

CASF™ — AI Safety Certified

Professional Level — COMING Q3 2026

Scenario-based exam + hands-on labs
Red teaming • Eval harness • Guardrails implementation
For: Security, ML, Risk, Compliance leads
Prereq: Foundations + 1yr experience or labs

$799 early (TBD) — coming later

Coming later

🔒

CASF™ — AI Safety Certified

Expert / Lead Level — COMING Q1 2027 — ANAB Track

Audit & Lead — ISO/IEC 42001 audit, NIST AI RMF 2.0
Case study + panel review
For: Auditors, Consultants, CISOs
Prereq: Professional + 2-3yr + case study

$1,499 + $199/yr — ANAB ISO 17024 track

Coming later

Foundations = understand risks & write policy. Professional = implement controls. Expert = audit & lead programs. Start at Foundations — no prerequisites.

BUILT ON FRAMEWORKS:NIST AI RMFISO 42001EU AI ActOWASP Top 10 for LLMs

How can I be AI safety certified?

Existing certs are either expensive, enterprise-security-only, or free completion certs with no exam and no verification.

TAISE — $795, 20–30h

Enterprise AI security. Heavy, pricey, security-first.

CASO — $2,199

Executive program. Great, but not for everyone.

BlueDot / CAIS — pay-what-you-want

Research / fellowship on-ramp. ~30h, apply, completion cert — not identity-verified, not workplace risk management.

See full pricing comparison

Why TAISE, CASO, SecAI+ aren't enough

TAISECASF™

Price$795$199

Time20–30 hours12 hours

FocusEnterprise securitySafety for builders

LanguageTechnicalPlain English

PrereqsIT backgroundNone

ExamYesIdentity-verified • 2 attempts

BadgePDFVerifiable QR + LinkedIn badge

AudienceSecurity prosPM, legal, ops, eng, founder

Save $596 vs TAISE • Save $2,000 vs CASO • Finish in a weekend

Security

Protect the model from attackers

Security = protect the model from attackers (prompt injection, breaches, access controls). Today's paid certs speak governance frameworks, DSPM, AISPM, and Zero Trust. Useful for CISOs. Wrong track for builders.

Safety

Protect people from the model

Safety = protect people, clients, and companies from what the model does (bias, hallucinations, leaking PII, over-reliance) — and a working AI Safety Policy you can take into a company, school, or agency.

PRACTICAL OUTCOME

Learn to write AI Safety Policies for organizations

The credential is for the job you already have. The artifact is a policy deployers, companies, schools, and agencies can actually use.

Companies and deployers

Write an AI Safety Policy that maps to NIST AI RMF and EU AI Act deployer duties — not a security checklist dressed up as safety.

Schools and agencies

Produce a one-pager leadership can actually adopt: what you will not deploy, who owns incidents, and how you review models.

The job you already have

Leave with a credential for the job you already have and a policy artifact you can show hiring managers, boards, or a compliance review.

6 MODULES • 12 HOURS • QUIZZES

Foundations Curriculum (12h)

M0Free now

Why AI Safety Is Not Security

Safety = protect people from the model. Security = protect the model from attackers. Write your first AI Safety Policy in 90 minutes.

M1Unlock $199

How AI Can Fail You

How AI systems fail in the real world, and how to start a risk register.

M2Unlock $199

Truth & Bias

Add Truth and Bias sections to the policy.

M3Unlock $199

Privacy & Confidentiality

Allowed tools list and data rules for what never goes in a prompt.

M4Unlock $199

Accountability & Law

EU AI Act and NIST AI RMF in plain English.

M5Unlock $199

Ship It: Final Policy

Ship the 5-page final AI Safety Policy.

Start free M0 Read the full curriculum

How certification works

0112h self-paced

Learn

6 modules, video + reading + quizzes. Plain English. Finish in a weekend. Lifetime access to updates.

6 modules

Video + reading

10Q quiz each

0260 questions • 60 min

Exam

Identity-verified online. 70% to pass. 2 attempts included. Instant results. No travel needed.

60 questions

70% pass

2 attempts

03Verifiable badge

Certified

Verifiable QR + LinkedIn badge at https://www.getaisafetycertified.com/verify/{id}. 2-year validity, $199 renewal, plus a policy you can take into an org. Foundations has no prerequisites. Professional and Expert stay later.

LinkedIn badge

QR verification

AI Safety Policy

PARTNERSHIPS

Built to partner with universities, research orgs, and standards bodies

We are seeking founding academic and institutional partners. No fake logos: we are not affiliated with TAISE, Northeastern, CSA, BlueDot, or CAIS. The path is a future university / research co-brand, not a claim we already have one.

Universities

A 12-hour, exam-backed foundations course you can offer as a credited module, continuing-ed certificate, or co-branded cohort — without pretending we already sit in your catalog.

Research orgs and standards bodies

A literacy layer that points at NIST AI RMF, ISO 42001, and the EU AI Act, then asks people to write a real policy. Built to partner, not to impersonate your seal.

Companies and chapters

Team playbooks for deployers and campus chapters that want a shared language for AI Safety Policy — career credential plus the artifact leadership can adopt.

Explore partnerships

Pricing

Foundations is $199. Organizations buy N × $199 seats with a private /org URL. The $2,999 Teams pack stays waitlist.

FOUNDATIONS

$199

vs $795 TAISE • vs $2,199 CASO

12-hour course (6 modules)

Identity-verified exam (60Q, 70% pass, 2 attempts)

Verifiable QR + LinkedIn badge at https://www.getaisafetycertified.com/verify/{id}

LinkedIn badge (1-click add)

AI Safety Policy you can take into an org

2-year validity + free updates

Start free M0 Unlock Foundations — $199

Free M0, then $199. Professional and Expert open later.

ORGANIZATIONS

N × $199

Unique /org URL for admins. Invite people. Employees keep /dashboard and their own certificates.

Create an organization

TEAMS

$2,999/ 20 seats

Later SKU. EU AI Act deployer policy pack. Not for sale yet.

Coming after Foundations

Foundations price $199 for every candidate or org seat. Payment unlocks M1–M5. The $2,999 Teams pack is not for sale yet.