National Cybersecurity
Association of Ukraine

A fake video or audio clip of a company executive: what to do.

Guide G11 · Published

For companies whose executive suddenly “appears” in a video or audio clip they never recorded: an instruction to staff, a statement to the public, or a message to investors. The order of work: verify, stop the money, keep evidence, say one short thing in public, and tell those who can help.

Who this guide is for and when to use it

This guide is for small and medium-sized companies and organisations where instructions about money, data or public statements are given in an executive’s name. Use it when you have found or received a video, voice message or call recording in which an executive appears to order or announce something and you are not sure it is genuine. The FBI warns that criminals generate videos for real-time video chats with alleged company executives, and use AI-generated audio to elicit payments 1.

If anyone is in danger or a crime is happening right now, call 102: the Cyber Police give this number for cases needing an immediate response 10.

The first hour: verify, stop the money, keep evidence

  1. 01Do not act on the instruction in the video or recording until it is confirmed. Reach the executive through a different channel from the one the request came on: dial a number you already know yourself, not the one in the message. The FBI advises verifying a caller by hanging up and calling a number you found independently 1 2.
  2. 02Stop payments, changes of bank details and access grants linked to the request. If money has already gone, contact your bank at once and ask whether the payment can be recalled.
  3. 03Tell the staff concerned (accounts, finance, IT, the executive’s office) that the instruction is unconfirmed and must not be carried out. Share what you learn only through verified internal channels.
  4. 04Do not forward or repost the fake, even “to warn people”.

Before the material disappears, preserve the evidence:

  1. 01Save the original file or the exact link as you received it. Do not edit, convert or compress it.
  2. 02Record the URL of every copy, the account name, the profile URL and the post ID. Give the date and time of each sighting with its time zone.
  3. 03Screen-record the page so the URL, account and date are visible, and add screenshots. Save emails as .eml files 9.
  4. 04Compute a hash (for example SHA-256) of every file and write it in the log. A hash is a fixed-length string produced by a hash function 12; it lets you show later that the file was not altered.
  5. 05Keep a log of who did what and when. Keep two copies, one of them offline.

A short holding statement while you check

Publish one short statement, only on your own verified channels, and do not embed the fake in it. Say no more than you know: do not call the material a “deepfake”, and do not cite an AI detector’s verdict, until you can back it up. One named person signs off the statement.

  1. 01What is known: a video or recording is circulating in which an officer of the company appears to speak.
  2. 02The company’s position: it did not make or authorise the statement (write this only after the internal check in step one).
  3. 03What others should do: not act on instructions in the material, not transfer money and not share data.
  4. 04Where the truth is: the address of your page of official channels and a contact for questions.

Separately, without linking to the fake, tell key partners and your bank how you normally contact them. The FTC advises a company to tell its staff in advance how it will contact them, so that they ignore messages that do not match 3.

Report to the platforms

Use each platform’s own reporting route. Start where the fake spreads most. Platforms change their forms often, so open the current help page before you report.

  1. 01YouTube, a channel or account posing as a person or company: use the impersonation report. YouTube does not allow unauthorised impersonation of a person, entity or channel, including AI copies of a voice or likeness, and it also enforces trademark holders’ rights 5.
  2. 02YouTube, a video with a generated likeness of your executive: a privacy complaint generally has to come from the person whose likeness was altered. YouTube does not accept complaints filed for colleagues or companies, so the executive in the video files it personally 6. You need the video URL and timestamps 6.
  3. 03Telegram: use the Report button on public content. For takedown you can also email abuse@telegram.org with the links. If someone is pretending to be you, its FAQ points to @NoToScam. Telegram does not process private chats and groups 7.
  4. 04TikTok: its rules require creators to label realistic AI content and allow unlabelled content to be removed, restricted or labelled, and they treat impersonation as deceptive behaviour 8. Cite this in an in-app report.
  5. 05Facebook, Instagram, X and others: open the platform’s current help centre and look for the impersonation report or the manipulated-media report. We could not verify their current steps, so we do not describe them here.
  6. 06Keep a copy of every report and every reply. Expect delays and a possible refusal: the platform makes the final decision.

Report to the authorities in Ukraine

  1. 01Cyber Police: the “Report a cyber incident” form is for phishing or malicious material (describe the event, attach a file up to 20 MB or an email as .eml). For other matters use cyberpolice.gov.ua/declare 9 10. Attach your evidence log.
  2. 02Cyber Police helpline: 0800 50 51 70 (Mon–Thu 9:00–18:00, Fri 9:00–17:00). The contacts page labels it for enquiries about e-appeals sent through the feedback form 10.
  3. 03CERT-UA: if the fake comes with a technical incident (a hacked email or account, malware, an intrusion into your network), write to incidents@cert.gov.ua or call +38 (044) 281-88-25 (Mon–Thu 8:00–17:00, Fri 8:00–15:45; at weekends and on public holidays, around the clock: +38 (044) 281-88-01) 15. Isolate the affected network first and have logs and suspicious files ready. CERT-UA works with technical data and is not a punitive body 11.
  4. 04Tell your bank, insurer and lawyer, and decide with counsel what to disclose and to whom.

Prepare in advance: agreements and account protection

  1. 01A second channel. Agree that any instruction about money, data or access received by voice, video or chat is confirmed by a callback to a known number or another channel fixed in advance 1 2.
  2. 02A codeword. The FBI suggests families agree a secret word or phrase to confirm identity 1. A company can adapt the idea for executives and the finance team (our adaptation, not FBI advice). Do not write the word in chat or email, and change it after use.
  3. 03Payment rules. No payment or change of bank details on a voice or video instruction alone, even one “from the director”, without confirmation by two people. Define the urgent-case procedure in advance instead of allowing exceptions.
  4. 04Who speaks for the company. Name a spokesperson and a deputy, prepare a template for the short statement and a page listing your official channels. Tell staff which channels executives do and never use to give instructions 3.
  5. 05Less material to fake. Limit public recordings of executives’ voice and image where you can 1, and review what your website and social media reveal about them and their assistants: that information makes scam messages more convincing 4.

Start protecting executives’ and assistants’ accounts with email: a compromised mailbox can open the door to other accounts, so turn on a second login step there, and on banking and social media 13. Prefer passkeys: each site gets its own, so a device cannot be tricked into logging in to a fake website 14. Keep your recovery options backed up separately 14. More in the guide “Protecting your accounts and digital identity”.

Limits

This is a guide, not legal advice and not a service. NAK receives no material from you, checks no files, removes no content, investigates no cases and does not respond to incidents. Ask a lawyer about liability, contracts, insurance and securing evidence for court. NAK does not claim that any tool can declare content real or fake. Platform steps change; this text was last checked on 24 Oct 2026. The sources used are listed below.

Sources

  1. 1FBI Internet Crime Complaint Center. Criminals use generative AI to facilitate financial fraud (PSA I-120324) · accessed 2026-10-24
  2. 2FBI Internet Crime Complaint Center. Senior US officials impersonated in malicious messaging campaign (PSA I-051525) · accessed 2026-10-24
  3. 3US Federal Trade Commission. Bringing new employees onboard? Help them avoid impersonator scams · accessed 2026-10-24
  4. 4UK National Cyber Security Centre (NCSC). Phishing attacks: defending your organisation · accessed 2026-10-24
  5. 5YouTube Help. Impersonation policy and how to report an impersonating channel · accessed 2026-10-24
  6. 6YouTube Help. Privacy complaint process (AI-altered or synthetic content) · accessed 2026-10-24
  7. 7Telegram. Telegram FAQ: reporting and abuse · accessed 2026-10-24
  8. 8TikTok. Community Guidelines: integrity and authenticity · accessed 2026-10-24
  9. 9Cyber Police Department, National Police of Ukraine. Cyber Police: incident form and appeals · accessed 2026-10-24
  10. 10Cyber Police Department, National Police of Ukraine. Cyber Police: contacts · accessed 2026-10-24
  11. 11Cabinet of Ministers of Ukraine. How CERT-UA responds to cyber incidents, from report to recovery · accessed 2026-10-24
  12. 12NIST CSRC. Glossary: hash value · accessed 2026-10-24
  13. 13CISA, Secure Our World. Turn on multifactor authentication · accessed 2026-10-24
  14. 14UK National Cyber Security Centre (NCSC). Trust the tech: using password managers and passkeys to help you stay secure online · accessed 2026-10-24
  15. 15CERT-UA. Contacts: report an incident · accessed 2026-10-24