---
title: Family office cybersecurity, now that AI is part of the attack surface
description: Family office cybersecurity now includes AI: the 2024 to 2026 survey numbers, how AI widens the family's attack surface, and what to ask of every AI vendor.
canonical: https://privatesuperintelligence.si/blog/family-office-cybersecurity-and-ai/
last-updated: 2026-10-15
---

# Family office cybersecurity, now that AI is part of the attack surface

Family office cybersecurity covers the family's money, documents, devices and people, and in 2026 it has to cover AI as well. Attackers use AI to write phishing emails and fake a principal's voice, staff paste family material into AI tools nobody approved, and AI assistants connected to email can act on what they read. The survey numbers below show where offices stand, and the sections after them show where AI changes the picture.

## What the surveys say about family office cybersecurity

Deloitte's Family Office Cybersecurity Report 2024 surveyed 354 single family offices between September and December 2023. Two newer reports from 2026 show the concern climbing.

| Finding | Figure | Source |
| --- | --- | --- |
| Family offices hit by a cyberattack in the previous 12 to 24 months | 43% globally, 57% in North America | Deloitte 2024 |
| Offices with more than US$1 billion AUM reporting an attack | 62%, against 38% for smaller offices | Deloitte 2024 |
| Attacked offices that received phishing emails | 93% (malware 35%, social engineering 23%, third-party risk 15%) | Deloitte 2024 |
| Offices with no cyber incident response plan | 31% | Deloitte 2024 |
| Offices with no disaster recovery plan | 50% | Deloitte 2024 |
| Offices that have not adopted "know your vendor" protocols | 68% | Deloitte 2024 |
| Offices naming cybersecurity and data breaches the top near-term operational concern | 59%, up from 16% the year before | RBC and Campden Wealth, September 2026 |
| Offices citing cybersecurity as their greatest service need | 32% | J.P. Morgan Private Bank, February 2026 |

The basics are mostly in place. Deloitte found that 85% of offices use strong passwords or multi-factor authentication and 72% back up their data. The gaps sit further along: only 58% train staff on cybersecurity, and Deloitte defines "know your vendor" as reviewing a vendor's security before you sign, for example by requesting security audit reports. That gap matters more each time the office adds a new AI tool, because every AI vendor is a third party holding family information.

## How AI widens the family's attack surface

AI enters a family office's risk picture in three places.

**Attackers use it.** IBM's 2025 Cost of a Data Breach Report, which studied breaches at 600 organizations, found that 16% of breaches involved attackers using AI tools, most often for phishing or deepfake impersonation. The FBI's Internet Crime Complaint Center (IC3) received 22,364 complaints in 2025 that referenced AI, with losses of $893,346,472. A well-written email in the principal's style used to take an attacker some effort. A language model now writes it in seconds, and a cloned voice can follow up by phone. The [deepfake wire fraud guide](/blog/deepfake-wire-fraud-family-offices/) covers that threat and the callback rules that stop it.

**Staff use it without telling anyone.** The controller drafts a capital account summary in a personal chatbot account. The executive assistant pastes the principal's travel plans into a free tool to build an itinerary. Each paste creates a copy of family information under terms the office never reviewed.

**The office connects it to everything.** An AI assistant that reads the principal's inbox, opens attachments and drafts replies has the same reach as a trusted employee. If it can also send email or move files, a hostile message can try to instruct it.

## Shadow AI: the copy of family data you don't know about

IBM uses "shadow AI" for staff using AI tools the organization hasn't approved. Its 2025 figures:

- One in five organizations reported a breach due to shadow AI.
- Organizations with high levels of shadow AI saw breach costs $670,000 higher on average than those with low levels or none.
- Only 37% have policies to manage AI or detect shadow AI.
- 63% of breached organizations either have no AI governance policy or are still developing one.

These figures cover organizations of every kind, and IBM did not break out family offices. The pattern still fits a small office: a lean team, many outside advisers, and no IT department watching which websites staff paste into. The fix starts with a written rule on which AI accounts staff may use for family work. Our [one-page AI policy for a single-family office](/blog/family-office-ai-policy/) gives you a template.

## AI tools that read email and take actions

The OWASP Top 10 for LLM Applications (2025) lists the risks that matter once an AI tool touches your inbox. LLM01 is prompt injection: text inside an email, document or web page that changes what the model does. LLM02 is sensitive information disclosure, and LLM06 is excessive agency, meaning a system holds more permissions than its job needs.

For a family office, the practical test is simple. Ask what the assistant can do without a person approving it. Drafting a reply is low risk. Sending it, forwarding attachments, or changing bank details in a bill pay system needs a human in the loop. IBM found that 13% of organizations reported breaches of AI models or applications, and 97% of those lacked proper AI access controls.

## The controls to put in place first

Deloitte's report closes with ten steps for family offices. Several of them apply to AI.

| Deloitte step | What it means once AI is in the office |
| --- | --- |
| Strong identity and access management, with access by role | Each AI tool gets only the mailboxes, folders and systems its role needs |
| A third-party risk management program that covers cyber risk | Every AI vendor goes through the same review as your custodian or IT provider |
| Basic training for all employees on phishing and malware | Training now includes cloned voices and deepfake video calls |
| A cyber incident response plan, exercised with key stakeholders | The plan names who to call if an AI tool exposes family data |
| A cybersecurity maturity assessment every one to two years | The assessment lists every AI tool in use, approved or not |

For the vendor step, use our [AI vendor due diligence checklist](/blog/ai-vendor-due-diligence-checklist/). It covers the five questions to ask before any AI tool reads family email: training, retention, processing location, permissions, and exit.

## For advisers who refer family offices

If you are a lawyer, accountant or banker asked to recommend family office cybersecurity firms, the Deloitte steps give you a neutral checklist to hand over. Ask any firm, and any AI vendor, to show how it covers identity and access, third-party risk, staff training and incident response. Ask for the answers in writing, so you can compare firms side by side.

## Private SuperIntelligence

Private SuperIntelligence from Mitosis Labs is an AI concierge for enterprises, family offices and private clients: the most capable AI, trusted with everything. It works across the family's own email, documents and data, answers with citations, and acts within permissions the office sets. Each deployment is built for one client, and a client deployment sends no telemetry back to Mitosis Labs. See [AI for family offices](/family-offices/) or read the [private AI guide for family offices](/blog/private-ai-for-family-offices/).

## Frequently asked questions

### How common are cyberattacks on family offices?

Deloitte's 2024 report found that 43% of family offices worldwide had a cyberattack in the previous 12 to 24 months. The figure was 57% in North America and 62% for offices with more than US$1 billion in assets.

### What is the most common cyberattack on a family office?

Phishing. Of the offices Deloitte surveyed that had been attacked, 93% had received phishing emails, and 35% had experienced malware.

### Does AI make family offices less secure?

AI adds risk in three places: attackers using it for phishing and deepfakes, staff using unapproved tools, and AI assistants with too many permissions. Each one has a known control, from callback rules to an approved-tools list to role-based access.

### What is shadow AI?

Shadow AI is staff using AI tools the organization hasn't approved. IBM's 2025 report found that one in five organizations had a breach due to shadow AI.

### What should a family office ask an AI vendor?

Ask whether it trains on your data, how long it keeps it, where processing happens, what the AI can do without approval, and how you get everything deleted when you leave. The [AI vendor due diligence checklist](/blog/ai-vendor-due-diligence-checklist/) sets these out as a table.


## Related reading

- [Private AI for family offices](/blog/private-ai-for-family-offices/)
- [Consolidated reporting across custodians, without the Excel export](/blog/family-office-consolidated-reporting/)
- [Tracking K-1s across 30 entities](/blog/k-1-tracking-across-entities/)
- [Capital call notices: catching every one, across every trust and LLC](/blog/capital-call-tracking/)
- [The family office technology stack](/blog/family-office-technology-stack/)
- [Family office operations, and where AI changes the staff math](/blog/family-office-operations-and-ai/)
- [What is private superintelligence?](/blog/what-is-private-superintelligence/)

## Sources

- [The Family Office Cybersecurity Report 2024 (Deloitte, 2024)](https://www.deloitte.com/mt/en/services/deloitte-private/research/family-office-cybersecurity-report.html)
- [The Family Office Cybersecurity Report 2024, full report PDF (Deloitte, 2024)](https://www.deloitte.com/content/dam/assets-shared/docs/services/deloitte-private/2024/family-office-cybersecurity-report-2024.pdf)
- [2026 RBC and Campden Wealth Report: North American family offices (PR Newswire, September 29, 2026)](https://www.prnewswire.com/news-releases/2026-rbc-and-campden-wealth-report-north-american-family-offices-shift-from-caution-to-optimism-amid-great-wealth-transfer-ai-acceleration-302892391.html)
- [J.P. Morgan Private Bank Releases 2026 Global Family Office Report (PR Newswire, February 2, 2026)](https://www.prnewswire.com/news-releases/jp-morgan-private-bank-releases-2026-global-family-office-report-302676012.html)
- [IBM Report: 13% of organizations reported breaches of AI models or applications (IBM Newsroom, July 30, 2025)](https://newsroom.ibm.com/2025-07-30-ibm-report-13-of-organizations-reported-breaches-of-ai-models-or-applications,-97-of-which-reported-lacking-proper-ai-access-controls)
- [2025 IC3 Annual Report (FBI Internet Crime Complaint Center, 2026)](https://www.ic3.gov/AnnualReport/Reports/2025_IC3Report.pdf)
- [OWASP Top 10 for LLM Applications 2025 (OWASP GenAI Security Project, accessed 2026-10-15)](https://genai.owasp.org/llm-top-10/)