Canada's Deepfakes Legislation Update: MPs Amend Bill to Criminalize Sexual Deepfakes Including 'Nearly Nude' Images

Canadian lawmakers have taken a significant step forward in combating the growing threat of AI-generated abuse by amending a key bill to explicitly criminalize the creation and distribution of sexual deepfakes, including those depicting individuals in 'nearly nude' states. This development, announced following a House of Commons justice committee meeting, expands protections for victims who have long faced devastating psychological and reputational harm from non-consensual synthetic imagery.

The amendment to Bill C-16, known as the Protecting Victims Act, addresses a critical gap in existing laws. Previously, Canada's Criminal Code section 162.1 prohibited the non-consensual sharing of intimate images—defined as those showing nudity or sexual activity—but it fell short against advanced AI tools that produce hyper-realistic fakes without requiring actual photos. Now, with 'nearly nude' included, the law aims to cover a broader range of exploitative content that humiliates and silences primarily women and girls.

This move reflects mounting public pressure amid rising reports of deepfake incidents. In early 2026, scandals involving AI chatbots generating explicit images spotlighted the urgency, prompting calls for swift action from advocates, tech experts, and everyday citizens alike.

Bill C-16 represents a comprehensive overhaul of criminal justice measures focused on protecting vulnerable groups. Introduced in December 2025 by Justice Minister Sean Fraser, it tackles multiple issues including femicide—murders motivated by misogyny or control—and coercive control in relationships. The deepfakes provision, however, has emerged as one of its most discussed elements.

Under the amended bill, an 'intimate image' now encompasses AI-generated visuals that realistically depict a person as nude, nearly nude, exposing genitals, anal region, or breasts, or engaged in explicit sexual acts, all without consent. Offenders face up to five years in prison for distribution and two years for possession or creation with intent to share. This builds on provincial civil remedies, like British Columbia's Intimate Images Protection Act, which allows victims to sue for damages and takedown orders.

The legislation defines deepfakes as synthetic media created using artificial intelligence to manipulate or fabricate a person's likeness in intimate contexts. This step-by-step process typically involves feeding public photos into generative AI models, which then output altered videos or images indistinguishable from reality to the untrained eye.

The pivotal change came during committee deliberations when Liberal and Conservative MPs united to broaden the scope beyond full nudity. Proponents argued that many deepfakes feature victims in revealing but not fully explicit attire, inflicting similar trauma through objectification and slut-shaming.

Bloc Québécois MP Rhéal Fortin voiced the sole dissent, warning that 'nearly nude' lacks precise boundaries. 'What constitutes nearly nude? A bikini? Lingerie? This vagueness could lead to overreach or inconsistent enforcement,' he cautioned, urging clearer guidelines to protect artistic expression or parody. Despite this, the amendment passed handily, signaling cross-party consensus on prioritizing victim safeguards.

Legal scholars note this mirrors evolutions in other jurisdictions. For instance, Australia's recent laws and U.S. state statutes have adopted similar expansive language to close loopholes exploited by perpetrators.

Deepfake sexual abuse has surged alongside accessible AI tools. Globally, over 95% of deepfakes target women, often celebrities or ex-partners, but ordinary Canadians are increasingly victimized. In Canada, financial losses from deepfake scams alone topped $77 million in 2025, with intimate image cases underreported due to stigma.

Real-world examples underscore the crisis. In January 2026, images generated by xAI's Grok chatbot proliferated on social media, depicting female public figures in compromising poses. Victims reported severe anxiety, job loss, and social isolation. Sextortion schemes, where perpetrators demand payment to halt distribution, have spiked, with the Canadian Anti-Fraud Centre noting a 30% rise in related complaints.

Statistics paint a grim picture: A 2026 survey found 88% of Canadians worry about AI misinformation, including deepfakes, eroding trust in digital media. Women aged 18-34 report the highest exposure, with one in five encountering manipulated content aimed at humiliation.

Advocacy groups like the Canadian Women's Foundation hail the bill as 'long overdue.' Executive director Pamela Cross emphasized, 'These images weaponize technology against women, perpetuating gender-based violence online. Criminalization sends a strong deterrent message.'

Tech ethicists offer mixed views. University of Toronto professor Katie Szilagyi praises the intent but flags enforcement hurdles: 'Attribution is tough when content spreads virally from anonymous accounts. We need better platform accountability.' Privacy Commissioner Daniel Therrien has launched probes into social media giants like X (formerly Twitter) for failing to curb deepfake proliferation.

Public sentiment leans supportive, with polls showing 75% favoring stricter AI regulations. Yet, free speech advocates caution against chilling satire, proposing narrow exceptions for journalistic or artistic use.

Implementing the law won't be straightforward. Proving intent and non-consent requires forensic analysis, as deepfakes evade traditional detection. Tools like Microsoft's Video Authenticator or Hive Moderation use AI to spot anomalies in lighting, facial movements, and pixels, achieving 90% accuracy in controlled tests.Microsoft's detection tool exemplifies progress, but widespread adoption lags.

Police face resource strains; RCMP cyber units are overwhelmed, handling 50,000+ complaints annually. International cooperation is vital, as servers often host content abroad. Experts recommend mandatory watermarking for AI outputs and faster takedown protocols, akin to the EU's AI Act.

• Trace origins via blockchain metadata embedded in AI generators.
• Train officers in deepfake forensics.
• Partner with platforms for proactive scanning.

Canada's approach aligns with a worldwide push. The U.S. TAKE IT DOWN Act mandates platforms remove non-consensual intimate images within 48 hours. UK's Online Safety Act criminalizes sharing, with fines up to £18 million. Australia amended laws in 2024 to ban creation, imposing two-year sentences.Justice Canada's Bill C-16 summary

Yet gaps persist. Unlike the EU, Canada lacks comprehensive AI risk classification, leaving non-sexual deepfakes (e.g., political misinformation) unregulated. This positions Bill C-16 as a targeted win but part of a larger needed framework.

Beyond individuals, deepfakes erode societal trust. In elections, fabricated videos could sway voters; a 2025 study found 25% of Canadians exposed to political fakes. Workplaces suffer from blackmail, and schools grapple with teen sextortion.

Positive shifts include industry self-regulation. Adobe and OpenAI now flag synthetic content, while startups develop victim support apps for rapid reporting. Education campaigns, like those from MediaSmarts, teach digital literacy to spot fakes.

Photo by Andy Holmes on Unsplash

With committee approval, Bill C-16 heads to second reading, potentially passing by summer 2026. Amendments may refine definitions, balancing protection and rights. Long-term, integration with Bill C-27's Artificial Intelligence and Data Act could mandate AI safety standards.

For victims, resources abound: Cybertip.ca for reporting, provincial hotlines for support. Tech firms must innovate faster than abusers, ensuring AI empowers rather than exploits.

This amendment marks progress in humanity's race against rogue AI, reminding us that law must evolve with technology to preserve dignity in the digital age.