This bill amends Nova Scotia’s Intimate Images and Cyber‑protection Act to bring artificially generated and materially altered sexual images within the statute’s protections and remedies. It inserts a three‑part definition of “fake intimate image,” creates a statutory duty for anyone who previously distributed an intimate image to “make every reasonable effort” to render it unavailable after the depicted person withdraws consent, and establishes civil liability for failure to do so.
The changes also lower procedural barriers to court relief by allowing applications even when the person in an image is not identifiable to third parties, and by placing the burden on respondents in personal intimate‑image cases to prove the applicant lacked a reasonable expectation of privacy. The bill therefore compresses legal thresholds that previously limited civil remedies and transfers new operational and legal obligations to distributors and online intermediaries.
At a Glance
What It Does
The bill expands the Act’s definition of intimate images to explicitly include AI‑generated or altered depictions that reasonably appear to show someone nude or engaged in sexual activity. It requires distributors to take every reasonable step to remove images after consent is revoked and makes them civilly liable for resulting injury if they fail. It also permits court applications regardless of whether the depicted person is identifiable and shifts the evidentiary burden to respondents in personal‑image cases.
Who It Affects
Online platforms, hosting services and intermediaries that store or redistribute images; individuals who create or share intimate images (including influencers and private sharers); civil litigants and defence counsel working on privacy and reputation cases; and victims of non‑consensual or AI‑manufactured imagery seeking remedies.
Why It Matters
This is one of the first provincial statutes to treat AI‑generated sexually explicit images as squarely within intimate‑image law, giving victims a clearer civil pathway for takedown and damages. The bill reassigns legal risks — especially evidentiary burdens and takedown obligations — which will change how platforms and lawyers handle removal requests and litigation.
More articles like this one.
A weekly email with all the latest developments on this topic.
What This Bill Actually Does
The bill rewrites the Act’s definitions so that intimate images now include both traditional recordings made in private and images that are artificially produced or manipulated to appear sexual. For artificially produced material the text uses a three‑part test: the image must reasonably and convincingly depict nudity or explicit sex, be created or altered by software or similar means, and be such that it’s reasonable to suspect the portrayed person would not have agreed to the recording or distribution.
That language draws a line intended to capture realistic deepfakes while excluding obviously fake or parody content, but it does so by reference to standards open to interpretation.
On takedown, the bill imposes a duty on anyone who has distributed an intimate image with consent to act if consent is later revoked. The duty is to “make every reasonable effort” to render the image unavailable; it is not limited to a single platform or to content that clearly identifies the subject.
If a distributor fails to make those efforts, the statute creates civil liability for injury caused by the failure. The remedy is therefore civil — damages and court orders appear to be the enforcement path rather than criminal sanctions.The bill also alters procedural access to the Supreme Court of Nova Scotia by allowing applications for orders even when the individual in the image can’t be identified by third parties from the image itself.
That change is aimed at covering realistic deepfakes and other altered media that nonetheless harm reputation or privacy without necessarily carrying identifiable metadata or facial likenesses. Finally, in disputes over personal intimate images (the non‑AI category recorded in private), the bill places the burden on the respondent to prove the applicant lacked a reasonable expectation of privacy at the time the image was made or distributed — a reversal of the usual presumption and a significant evidentiary shift.Taken together, these amendments expand who can seek civil relief, narrow factual defenses available to distributors, and create operation‑level obligations for anyone who has shared intimate images.
While the text offers a clearer route to relief for victims of deepfakes and non‑consensual sharing, it also leaves several practical questions unanswered about what constitutes reasonable removal efforts, how cross‑border distribution will be handled, and how courts should assess expectations of privacy in edge cases.
The Five Things You Need to Know
The bill defines a “fake intimate image” using three criteria: a reasonably convincing sexual depiction, creation or alteration through software/AI, and a reasonable suspicion the depicted person would not consent.
It requires anyone who previously distributed an intimate image with consent to make “every reasonable effort” to make that image unavailable after consent is revoked and imposes civil liability for injuries caused by failing to do so.
The law allows a person to apply to the Supreme Court of Nova Scotia for orders related to non‑consensual distribution even when the person in the image cannot be identified by third parties.
For personal intimate images, the respondent must prove the applicant did not have a reasonable expectation of privacy when the recording was made or when it was distributed — shifting a key evidentiary burden onto defendants.
Amendments are civil in nature: they alter the Intimate Images and Cyber‑protection Act’s definitions, remedies, and procedural access rather than adding new criminal offences.
Section-by-Section Breakdown
Every bill we cover gets an analysis of its key sections.
Short title
A concise technical provision: the statute will be cited as the Distribution of Fake Intimate Images Act. This does not change substance, but signals the legislature’s focus on artificially generated images.
Adds 'fake intimate image' and 'personal intimate image' and expands 'intimate image'
The bill replaces the existing definition framework with two distinct categories. The fake‑image definition couples a subjective impact test ('reasonably convincing') with objective creation mechanics (software/AI or manipulation of authentic images) and a normative consent expectation (reasonable suspicion of lack of consent). The personal intimate‑image definition layers two privacy‑time tests: circumstances giving rise to a reasonable expectation of privacy when recorded, and retention of that expectation at distribution. Practically, those multi‑part definitions will be litigated: courts must interpret what ‘reasonably convincing’ looks like, when manipulation crosses the line into the statute, and how to time‑slice an applicant’s privacy expectation.
Duty to make images unavailable after consent is revoked; damages for failure
This section obliges anyone who distributed an intimate image with consent to take every reasonable step to remove access if the subject revokes consent. It attaches civil liability for injuries resulting from a failure to act. The statutory phrase 'every reasonable effort' is deliberately broad; compliance could require takedown notices, de‑indexing, contacting downstream hosts, or using technological measures. Because the bill leaves the exact contours undefined, parties will rely on case law and proportionality — such as the distributor’s control over the content, available resources, and the scale of distribution — to determine what counts as reasonable.
Court access regardless of identifiability
The amendment allows applications for orders under the Act even if the person depicted is not identifiable to third parties from the image. This explicitly captures anonymous or heavily altered deepfakes. The practical implication is lower procedural standing barriers for applicants, but courts will still assess whether the applicant can demonstrate distribution and harm. It also raises evidentiary questions about proving distribution when identifiers are absent and about the scope of injunctive relief in cases involving widely syndicated anonymous content.
Burden shift in personal intimate‑image claims
When the claim concerns a personal intimate image made or distributed without consent, the respondent must prove the applicant lacked a reasonable expectation of privacy at the relevant time. That reallocation of proof will reshape defence strategies: respondents will need documentary or contextual evidence (consent statements, publication settings, public posting that eliminated privacy expectations). For claimants it lowers the initial evidentiary bar but will likely encourage deeper factual discovery and contested hearings over contextual circumstances.
This bill is one of many.
Codify tracks hundreds of bills on this topic across all five countries.
Explore this topic in Codify Search →Who Benefits and Who Bears the Cost
Every bill creates winners and losers. Here's who stands to gain and who bears the cost.
Who Benefits
- People targeted by AI‑generated sexual images: they gain clearer statutory recognition that realistic deepfakes constitute intimate images and a civil pathway for takedown and damages.
- Victims of non‑consensual sharing: the takedown duty and lowered identifiability threshold make it easier to request removal and seek remedy even when images lack clear identifiers.
- Privacy and support NGOs: the statutory changes create clearer legal claims these organizations can use to advocate, assist with takedowns, and litigate on behalf of clients.
- Providers of takedown and compliance services: businesses that offer content‑removal, monitoring, and de‑indexing services will see increased demand as distributors seek to meet the 'every reasonable effort' standard.
Who Bears the Cost
- Online platforms and intermediaries: they face more takedown requests, potential follow‑on litigation, and operational costs to detect and remove AI‑generated images, especially where distributors lack direct control over downstream hosts.
- Individual distributors (including private sharers): people who once shared images with consent now risk civil liability if they cannot or do not adequately remove content after consent is withdrawn.
- Small hosting services and community platforms: complying with broad removal efforts and responding to contested claims may strain limited moderation resources and legal budgets.
- Courts and legal services: the Supreme Court will likely see more civil applications and contested hearings, increasing caseload and creating a need for evidentiary frameworks around deepfakes and privacy expectations.
Key Issues
The Core Tension
The bill balances the urgent need to protect people from realistic, non‑consensual sexual imagery against the risk of overbroad takedowns and burdens on intermediaries: it accelerates victims’ access to civil remedies and forces distributors to act quickly, but does so by leaving high‑stakes, discretionary standards to courts and imposing operational costs that may be hard to distribute fairly.
The bill resolves an important gap — recognizing realistic AI‑produced sexual images as the same kind of privacy harm as non‑consensual recordings — but it relies on several indeterminate standards. Terms such as 'reasonably convincing', 'reasonable suspicion' and 'every reasonable effort' are flexible by design, leaving courts to define them case by case.
That approach allows fact‑sensitive outcomes but creates uncertainty for platforms and distributors trying to design compliant procedures in advance.
Cross‑border enforcement and technical feasibility are other unresolved problems. Popular image hosting and social networks operate globally; a Nova Scotia civil order has limited reach against offshore hosts unless platforms cooperate voluntarily or take pre‑emptive measures.
The bill’s civil‑only enforcement model gives victims a path to damages and injunctive relief, but it may not deter sophisticated creators of deepfakes who operate anonymously. Finally, shifting the burden to respondents in personal‑image claims aids victims procedurally but risks chilling legitimate speech if takedown obligations are applied too aggressively without clear expedited processes for contesting wrongful removals.
There's more to this law than the bill.
Codify Laws traces every connection across the legislative lifecycle.