Fake corporate announcements: proving what the issuer published
A listed company speaks to the market through a few formal channels: releases sent through regulated information channels, analyst presentations, earnings calls, investor relations pages. Everything else is commentary. What has changed is the cost of forgery. A fake corporate announcement takes minutes to assemble, travels as a screenshot before anyone can check it, and sometimes arrives with the synthetic voice of the chief executive attached.
Markets move on the first version in circulation, not the most reliable one. No crisis protocol answers the question that follows: how does an issuer show, immediately and verifiably by third parties, what the authentic communication said? An issuer cannot win by chasing the fake. It wins by certifying at source, at publication, what it actually communicates, so that a verifiable official version always exists to set against anything circulating.
Why a fake corporate announcement moves the price before the denial
The price moves before the denial because verification always arrives after the buy order. In April 2013 the Associated Press Twitter account was compromised and used to report an explosion at the White House. The S&P 500 lost about 136.5 billion dollars of market capitalisation in three minutes, a figure attributed to Reuters and reported by CNBC, before the index recovered once the hack was confirmed. Machines widen the window: more than 66% of global equity volume is handled by algorithmic traders, according to Aite Group, and an automated system cannot tell an authentic release from a fabricated one. Research by Paolo Pellizzari of Ca' Foscari University of Venice with Sarah Mignot and Frank Westerhoff of Bamberg, published in 2024 in the Journal of Economics and Statistics, finds that false news decouples an asset's price from its fundamentals and produces endogenous chaotic oscillations that do not settle on their own after a correction.
The pattern varies little. Content originates outside the regulated channel and re-enters sideways, through accounts impersonating a press office or a cropped image stripped of any trace of where it came from. Whoever receives it has nothing to compare it against, because the original carries nothing that marks it as the original.
Synthetic voice and video: when a deepfake of the chief executive becomes market news
A deepfake of the chief executive works as market news because it moves the forgery from the document to the person, and audiences grant a voice and a face a credibility they never grant plain text. In July 2024 a Ferrari executive received WhatsApp messages from someone posing as chief executive Benedetto Vigna, then a call in which a cloned version of Vigna's voice pressed for discretion over an imminent transaction. The attempt collapsed when the executive asked about a detail only the real Vigna could have known, as reported by Fortune. At the engineering group Arup, a finance employee in Hong Kong authorised fifteen transfers worth about 25 million dollars after a video call in which every other participant was a synthetic reconstruction, according to CNN. Anyone with a few minutes of public audio can experiment with how synthetic voice and video work, and a chief executive's quarterly earnings call is training material already published.
Neither episode began as an attack on a share price, but the same technique pointed at a results announcement produces one. A fake press release can be set against the original, if the original was certified. An audio file attributed to an earnings call that never happened has nothing to be set against. A deepfake CEO video aimed at the market is the voice cloning problem with a wider reach.
Market manipulation: when financial disinformation becomes market abuse
Information-based manipulation is the dissemination of information that gives, or is likely to give, false or misleading signals as to the supply of, demand for or price of a financial instrument. Regulation (EU) No 596/2014, the Market Abuse Regulation, treats it as one form of market manipulation. MAR applies whether the information travels through the media, including the internet, or by any other means, and it does not require the manipulation to have worked: what counts is whether the information was capable of distorting the price. A fabricated release attributed to a listed company falls inside the definition even if the share price never moved. The perimeter covers instruments admitted to trading on regulated markets, multilateral trading facilities and organised trading facilities, which brings growth-market and pre-IPO issuers into scope alongside blue chips. Financial disinformation becomes supervised conduct rather than a reputational nuisance.
The forms of market abuse
MAR defines three prohibited conducts, and together they mark the perimeter of market abuse. A false corporate announcement sits in the third, in its informational variant.
| Conduct | What it involves | Reference |
|---|---|---|
| Insider dealing | Dealing on non-public, price sensitive information | Reg. (EU) 596/2014, Art. 8 |
| Unlawful disclosure of inside information | Passing inside information to a third party outside normal duties | Reg. (EU) 596/2014, Art. 10 |
| Market manipulation | Distorting price formation through false information or fictitious transactions | Reg. (EU) 596/2014, Art. 12 |
The first two presuppose that the person holds price sensitive information before the market does. The third does not: anyone can commit it, including an outsider who only fabricates a communication in the issuer's name.
Information-based versus trade-based manipulation
The distinction turns on the instrument of the abuse, not its effect. Information-based manipulation works through content: false statements, rumours, fabricated documents, planted research. Trade-based manipulation works through orders, from wash trades to the pump and dump schemes familiar from small-cap markets. Supervisors reconstruct the second kind from order books and transaction reports they already hold. The first leaves its trace outside the trading infrastructure, in content that mutates as it circulates.
How supervisors respond
ESMA sets the common standard and national competent authorities enforce it. ESMA issues guidelines, technical standards and Q&As on MAR through its market integrity workstream. The competent authority in each member state can require documents from any person, take statements, carry out inspections and cooperate across borders. The regulation sets the sanctioning floor itself: for breaches of the insider dealing and market manipulation prohibitions, maximum administrative fines must reach at least 15 million euro, or 15% of annual turnover, for legal persons. For an issuer, a documentary request arrives sooner or later.
Why a denial is not evidence: the issuer's evidentiary asymmetry
A denial is an act of communication, not an item of evidence: it attests what the issuer states today, not what it published yesterday. It carries no trusted date and no integrity fingerprint tied to the original document, and it lands after the false content has circulated in places the correction never reaches. The asymmetry is structural. The forger only has to make the fake plausible, while the issuer has to reconstruct a technical fact after the event. According to an analysis published in the Capital Markets Law Journal in 2026, EU regulatory frameworks, MAR included, remain fragmented and reactive with respect to manipulation based on synthetic media, so protection arrives once the effect on the price has been produced. Nor does time repair the damage: the Bamberg and Ca' Foscari model finds that oscillations set off by false information capable of distorting the price keep running after the correction.
False content almost always circulates as an image, and a screenshot is not evidence merely because it exists. The same holds, roles reversed, for the copy an issuer pulls off its own website weeks later: the admissibility requirements for digital evidence look at provenance and integrity.
| Dimension | Denial after the fake | Certified at publication |
|---|---|---|
| Time to availability | Hours or days, once the stock has moved | Immediate: the document predates the event |
| Trusted date | Absent: the denial dates itself | Qualified timestamp tied to the release |
| Content integrity | Later edits cannot be excluded | Verifiable via cryptographic fingerprint |
| Third-party verifiability | Requires trust in the issuer | Checkable by analysts, supervisors, courts |
Regulated disclosure channels are not enough
Regulated information disclosure channels guarantee simultaneous, traceable distribution, and nothing past that. Regulatory news services put regulated information in front of the whole market at once and leave a record with the competent authority, but they do not follow the content downstream. Once the text is cropped into an image and pushed back out through a messaging group, it leaves the channel and loses the link to its origin. A corporate disinformation protection programme that stops at the boundary of the official channel covers the distribution and leaves the artefact undefended.
This is what certification at source means: not chasing the fake, but making the authentic provable at the moment it is published.
How do you prove what the issuer actually published?
TrueScreen certifies a press release at the moment the issuer publishes it, applying a qualified timestamp and an integrity fingerprint to the content as released. The qualified timestamp is issued by a third-party qualified QTSP integrated via TrueScreen's API, and it fixes a date that can be asserted against third parties. The cryptographic fingerprint makes any later change detectable, down to a single character, and an electronic seal compliant with the eIDAS Regulation binds both to the document. The same operation covers analyst presentations, earnings call recordings and the option to certify an investor relations page as it appeared at a given instant. On the scale of the problem there are estimates: the Sopra Steria study puts the global economic impact of disinformation in 2024 at 417 billion dollars, of which 393 billion in financial flows diverted by information manipulation and 11 billion tied to deepfakes and AI-enabled fraud.
At 8:14 on a Monday an account impersonating an issuer's press office posts a screenshot of a release announcing the withdrawal of full-year guidance. At 8:31 the stock falls in pre-market. At 9:05 investor relations drafts a denial, but the question coming from the supervisor and from analysts is different: what exactly was published on the website at 7:00, and how is that demonstrated. If every release has been sealed at publication, the answer is a file produced within minutes. If not, days of reconstruction across server logs and third-party caches begin.
Organisations use TrueScreen to seal analyst presentations, earnings call recordings and investor relations pages, so that every official version stays verifiable months later, with certified reports exportable as PDF and JSON. The technical anchors are the eIDAS 2 requirements for qualified timestamps and the forensic copy of the site, when a page has to be captured as it stood at a precise moment. Hostile content works the same way: acquiring it while it is still online is what makes it contestable later.
The conversation with analysts and authorities then changes subject: the issuer's document, with its date and its fingerprint, rather than the plausibility of the fake. That is certification at source applied to market communication, the job of the Data Authenticity Platform rather than of a detection tool.
