AI Scarlett Johansson Sex: Ethical & Legal Deepfakes

The advent of artificial intelligence (AI) has ushered in an era of unprecedented digital innovation, transforming industries from healthcare to entertainment. One of the most fascinating, yet profoundly controversial, applications of AI lies in its capacity to generate incredibly realistic synthetic media. This includes images, audio, and video that are virtually indistinguishable from genuine recordings. While some applications, like AI-driven animation or virtual assistants, offer immense positive potential, others raise serious ethical, legal, and societal concerns. The ability to manipulate or create digital content at this level, particularly when it pertains to individuals, demands a deep dive into the technology's implications. The core of this capability lies in advanced machine learning models, primarily Generative Adversarial Networks (GANs) and variational autoencoders (VAEs). GANs, for instance, involve two neural networks—a generator and a discriminator—pitted against each other. The generator creates new data (e.g., an image of a face), and the discriminator tries to determine if the data is real or fake. Through this adversarial process, the generator becomes incredibly adept at producing highly convincing synthetic content. VAEs, on the other hand, learn a compressed representation of input data and can then generate new data by sampling from this learned representation. These powerful tools, when combined with vast datasets, enable the creation of highly convincing digital fakes. The rapid progression of computational power and algorithmic sophistication has democratized access to these technologies. What was once the domain of specialized visual effects studios is now, to varying degrees, accessible to individuals with standard computing resources and publicly available software libraries. This accessibility, while promoting creativity and technological exploration, also introduces significant risks, particularly when used maliciously. The capacity to create digital representations of individuals without their consent, especially in sensitive contexts, represents a critical challenge to personal privacy and digital security in 2025.

The term "deepfake" itself is a portmanteau of "deep learning" and "fake," originating around 2017 to describe the use of deep learning techniques to synthesize or manipulate videos and other digital media to a high degree of realism. Initially gaining notoriety through online forums where users shared manipulated videos, the technology quickly moved into the mainstream consciousness, often associated with celebrity impersonation and non-consensual explicit content. Deepfakes operate by mapping an individual's facial expressions, movements, and voice onto another person's body or voice, or by creating an entirely new, fabricated scenario. The process typically involves training an AI model on a large dataset of images and videos of the target individual. The more data available, the more realistic and convincing the deepfake becomes. For public figures, whose likenesses are widely available online, this abundance of source material makes them particularly vulnerable targets for synthetic media manipulation. The sophistication of deepfake technology has evolved rapidly. Early deepfakes might have exhibited noticeable artifacts, blurring, or inconsistencies that betrayed their artificial nature. However, by 2025, advanced deepfake algorithms can produce content that is nearly indistinguishable from reality, even to the trained eye. This technological leap has profound implications, blurring the lines between what is real and what is fabricated, and challenging our ability to trust digital evidence. Beyond visual manipulation, AI can also generate realistic voices and even entire conversations, adding another layer of complexity to the landscape of synthetic media. This auditory aspect further complicates efforts to discern authentic content from fabricated ones, creating a powerful tool for misinformation and deceptive practices. The motivation behind creating deepfakes varies widely, from harmless parody and satirical commentary to malicious intent, such as defamation, fraud, political disinformation, or the creation of non-consensual explicit content. It is this latter category, particularly involving public figures, that has ignited widespread concern and prompted urgent calls for ethical guidelines and robust legal frameworks.

Scarlett Johansson, a globally recognized actress, has unfortunately become a prominent example of a public figure whose likeness has been exploited through non-consensual deepfake technology. Her high profile, extensive media presence, and widespread public recognition make her a frequent target for those seeking to create and disseminate fabricated explicit content. The phenomenon of "ai scarlett johansson sex" as a search term highlights a deeply troubling trend: the weaponization of AI to create and circulate intimate material without consent, specifically targeting celebrities. These deepfakes typically involve superimposing Johansson's face onto the bodies of individuals in explicit videos or images, or generating entirely new scenarios using her likeness. The intent behind such creations is often malicious, aiming to degrade, humiliate, or financially exploit the individual. The psychological and professional impact on victims can be severe, ranging from reputational damage and emotional distress to feelings of violation and a loss of control over their own digital identity. It's a digital form of harassment and exploitation that transcends traditional boundaries, reaching a global audience instantaneously. The targeting of a figure like Scarlett Johansson underscores a broader vulnerability that affects countless individuals, both public and private. While celebrities have the platform to speak out, the vast majority of deepfake victims are ordinary citizens, often women, who have fewer resources to combat the spread of such content or seek legal redress. The very nature of the internet – its decentralization and speed of dissemination – makes it incredibly challenging to remove deepfakes once they are online. This digital indelible ink means that once a deepfake is created and shared, it can persist indefinitely, resurfacing years later and continuously causing harm. Johansson herself has been vocal about the issue, highlighting the violation of privacy and the challenges of combating such sophisticated forms of digital abuse. Her experience serves as a stark reminder of the urgent need for comprehensive strategies to address the misuse of AI in creating non-consensual intimate imagery. The existence of "ai scarlett johansson sex" content isn't just about a single celebrity; it's a canary in the coal mine, signaling a fundamental threat to digital consent and personal sovereignty in the age of advanced AI. It forces a critical examination of what it means to control one's image and identity when AI can so effortlessly warp reality.

The creation and dissemination of "ai scarlett johansson sex" content, and deepfake pornography in general, presents a profound ethical quagmire. At its core, the issue strikes at fundamental principles of consent, autonomy, and human dignity. Unlike traditional forms of media manipulation, deepfakes can generate entirely new realities, placing individuals in situations they never consented to, often of an extremely intimate and humiliating nature. Violation of Consent and Autonomy: The most pressing ethical concern is the egregious violation of consent. Individuals depicted in non-consensual deepfakes have no control over the creation or distribution of these images. Their likeness is stolen, manipulated, and presented in a context that is entirely fabricated and deeply personal. This act strips individuals of their autonomy over their own bodies and digital identities, reducing them to objects of digital manipulation. It’s a profound invasion of privacy that goes beyond typical paparazzi intrusion, as it fabricates scenarios rather than merely capturing existing ones. Dignity and Reputation: For victims, the impact on their dignity and reputation can be devastating. Whether a public figure like Scarlett Johansson or a private individual, being associated with fabricated explicit content can lead to severe emotional distress, shame, and professional repercussions. The digital permanence of these fakes means that even if removed from one platform, they can resurface elsewhere, perpetually haunting victims and damaging their public and private lives. The pervasive nature of online content means that even if a deepfake is debunked, the initial exposure and its associated stigma can linger indefinitely. Misinformation and Trust Erosion: Beyond individual harm, the prevalence of deepfakes erodes trust in digital media and information. If sophisticated AI can fabricate such convincing videos, how can individuals, news organizations, or even legal systems discern truth from fiction? This trust deficit has far-reaching implications for journalism, political discourse, and the justice system, potentially fostering a climate of pervasive doubt where genuine evidence can be dismissed as fake and fabricated content accepted as real. This erosion of trust can destabilize democratic processes and undermine public discourse by making it difficult to establish shared facts. Gendered Harm: It is crucial to note that the vast majority of non-consensual deepfake pornography targets women. This phenomenon is a digital extension of gender-based violence and harassment, often motivated by misogyny and the desire to control, exploit, and degrade women. It weaponizes technology to perpetuate existing societal inequalities and power imbalances. The disproportionate targeting of women highlights the need for gender-sensitive approaches in policy and technological solutions. Responsibility of Developers and Platforms: The ethical burden also extends to the developers of AI technology and the platforms that host and disseminate deepfakes. While the technology itself is neutral, its potential for misuse demands a high degree of responsibility from its creators and those who provide the infrastructure for its distribution. Questions arise about the moral obligation to implement safeguards, develop detection tools, and enforce strict content moderation policies to prevent harm. The ethical responsibility isn't just about what is legal, but what is morally right, especially when the potential for severe harm is evident. Companies have a role in proactively designing technologies that are less susceptible to malicious use and in quickly responding to reports of harmful content. In 2025, the ethical landscape surrounding AI-generated explicit content remains a critical area of debate, pushing the boundaries of traditional legal and moral frameworks and necessitating a collective re-evaluation of digital rights and responsibilities.

The rapid advancement of deepfake technology, particularly its misuse in creating non-consensual explicit content like "ai scarlett johansson sex," has presented significant challenges to existing legal frameworks worldwide. Traditional laws designed to address defamation, copyright, or harassment often struggle to adequately capture the nuances and scale of harm caused by AI-generated media. However, by 2025, many jurisdictions have begun to implement specific legislation or adapt existing laws to combat the deepfake menace. Existing Legal Challenges: * Defamation: While deepfakes can certainly be defamatory, proving malicious intent and demonstrable harm can be complex, especially if the content is hosted anonymously or internationally. * Copyright Infringement: While the original source material might be copyrighted, the transformative nature of deepfakes complicates direct copyright claims on the generated content itself. The image of a person, unless specifically trademarked or protected under publicity rights, often isn't directly covered by copyright for simple facial appropriation. * Revenge Porn Laws: Many jurisdictions have enacted laws against the non-consensual sharing of intimate images (NCII), often referred to as "revenge porn" laws. The question arises whether deepfakes, being synthetic, fall under these definitions, which typically require an "actual" image or video of the victim. Some laws have been updated to explicitly include synthetically generated content. * Impersonation and Identity Theft: While deepfakes involve impersonation, they don't always involve financial fraud or direct identity theft in the traditional sense, making existing laws in these areas sometimes insufficient. Emerging Legal Frameworks and Responses (by 2025): * Specific Deepfake Legislation: Several countries and U.S. states have passed laws specifically targeting the malicious creation and distribution of deepfakes, particularly those of a sexual nature. These laws often make it a criminal offense to create or disseminate non-consensual synthetic intimate imagery. For example, some states in the US have introduced bills that allow victims to sue for damages. * Right of Publicity/Personality Rights: These laws, prevalent in some jurisdictions, grant individuals the exclusive right to control the commercial use of their name, image, likeness, and other aspects of their identity. Deepfake creation, especially if for commercial gain or widespread distribution, can be challenged under these rights. Scarlett Johansson, as a public figure, might have strong grounds for action under such laws. * Criminalization of Non-Consensual Synthetic Intimate Imagery: The trend is towards explicitly criminalizing the creation, possession, and distribution of deepfake pornography, treating it with the same gravity as non-consensual real imagery. Penalties can include fines and imprisonment. * Platform Accountability: Increasingly, legislation is being considered or enacted that places greater responsibility on social media platforms and content hosts to promptly remove deepfakes and implement robust detection and reporting mechanisms. The Digital Services Act (DSA) in the European Union, for example, puts significant obligations on large online platforms to mitigate risks, including those posed by deepfakes. * Misinformation and Election Integrity: Beyond explicit content, deepfake laws are also being developed to combat political deepfakes designed to spread misinformation and influence elections, often requiring disclaimers or prohibiting deceptive content during election periods. Despite these efforts, legal enforcement remains a significant challenge. The global nature of the internet means that deepfakes can be created in one country, hosted in another, and accessed worldwide, complicating jurisdictional issues. Identifying perpetrators, especially those operating anonymously, is also a major hurdle. However, the increasing legal pressure signals a global recognition of the severity of the deepfake threat and a growing commitment to protecting individuals from this form of digital abuse. Legal frameworks in 2025 are still catching up to the technology's pace, but the intent to protect digital consent and prevent harm is clear.

The battle against non-consensual synthetic media, including explicit deepfakes featuring individuals like Scarlett Johansson, is multifaceted, involving technological, legal, and societal interventions. It's a complex endeavor requiring collaboration across governments, tech companies, law enforcement, and civil society organizations. Technological Solutions: * Detection and Authentication: Significant research and development are underway to create sophisticated AI tools capable of detecting deepfakes. These tools look for subtle artifacts, inconsistencies, or statistical anomalies in synthetic content that human eyes might miss. Watermarking technologies are also being explored, where digital signatures are embedded into authentic media to verify its origin. However, this is an arms race: as detection methods improve, so do the methods of deepfake generation, making it an ongoing challenge. * Source Verification: Blockchain technology and cryptographic methods are being investigated to create immutable records of content origin, allowing users to verify if a piece of media has been altered or is entirely synthetic. This would establish a "chain of custody" for digital content. * Content Moderation AI: Platforms are increasingly deploying AI-powered content moderation systems to identify and remove deepfakes automatically. While these systems are improving, they are not foolproof and require human oversight, especially for nuanced or novel forms of manipulation. Legal and Policy Interventions: * Streamlined Reporting Mechanisms: Platforms are improving their reporting tools for victims of deepfakes, making it easier to flag and request removal of infringing content. * International Cooperation: Given the global nature of deepfakes, international cooperation is crucial for effective law enforcement and content removal across borders. Treaties and agreements are being explored to facilitate cross-jurisdictional action against perpetrators. * Victim Support: Providing psychological, legal, and technical support for deepfake victims is paramount. Organizations are emerging to help individuals navigate the emotional trauma, legal complexities, and practical steps of combating online abuse. Societal and Educational Approaches: * Media Literacy: A crucial long-term strategy is to improve public media literacy. Educating individuals on how to critically evaluate online content, understand the capabilities of AI manipulation, and recognize potential deepfakes is essential for building a more resilient digital society. This involves teaching people to question the source, look for tell-tale signs of manipulation, and be skeptical of emotionally charged content. * Ethical AI Development: Encouraging ethical guidelines and practices within the AI development community is vital. This includes prioritizing privacy-preserving AI, developing "red teaming" exercises to identify potential misuse, and integrating ethical considerations from the design phase of new technologies. Responsible AI development aims to build in safeguards from the ground up, rather than retrofitting them after harm has occurred. * Advocacy and Awareness Campaigns: Highlighting the harm caused by non-consensual synthetic media through public campaigns can raise awareness and reduce the demand for such content. The experiences of individuals like Scarlett Johansson can be powerful tools in these campaigns, giving a human face to the problem. * Industry Standards: Encouraging tech companies to adopt common industry standards for content authenticity and deepfake detection can create a more unified front against malicious actors. This might include shared databases of known deepfake patterns or collaborative research efforts. The fight against non-consensual synthetic media is an ongoing arms race. As technology advances, so do the methods of both creation and detection. A multi-pronged approach that combines advanced technological solutions, robust legal frameworks, and comprehensive societal education is necessary to mitigate the pervasive threat posed by deepfakes in 2025 and beyond.

The proliferation of advanced AI, capable of generating incredibly convincing synthetic media, poses a unique and escalating threat to the privacy and personal autonomy of both public figures and private citizens. The case of "ai scarlett johansson sex" exemplifies how readily a public figure's identity can be co-opted and exploited without consent, highlighting the urgent need for comprehensive protective measures in 2025. Challenges for Public Figures: Public figures, by definition, have their images, voices, and personal details widely accessible through media, interviews, and public appearances. This extensive digital footprint provides an abundant dataset for AI models, making them particularly vulnerable to deepfake creation. While they may have a platform to speak out, the sheer volume and speed of online dissemination mean that harmful deepfakes can spread globally before effective countermeasures can be deployed. Their public nature also means that such content can be more readily believed or amplified, causing greater reputational and emotional damage. Furthermore, the line between legitimate public interest and intrusive exploitation becomes increasingly blurred when AI can fabricate entirely new, intimate realities. Challenges for Individual Privacy: For private citizens, the threat is equally insidious, albeit often less visible. While they might not have a vast public image dataset, malicious actors can still leverage private photos, social media content, or even limited personal information to create deepfakes. The impact on an ordinary individual can be even more devastating, as they typically lack the legal resources, public platform, or support networks available to celebrities. The emotional trauma, social ostracization, and professional repercussions can be profound and long-lasting, often without effective recourse. The very notion of controlling one's own image and narrative in the digital age is under assault. Strategies for Protection in 2025: 1. Proactive Legal Frameworks: * "Right to be Forgotten" and Digital Erasure: Strengthening laws that allow individuals to request the removal of harmful or non-consensual content, including deepfakes, from online platforms and search engine results. * Stronger Privacy Regulations: Expanding the scope of privacy laws (like GDPR in Europe or evolving data privacy acts globally) to explicitly cover the unauthorized use of biometric data and digital likenesses for AI generation. * Civil Remedies: Establishing clear pathways for victims to sue creators and distributors of deepfakes for damages, including emotional distress, reputational harm, and economic loss. 2. Technological Safeguards and Authentication: * Authenticity Infrastructure: Developing industry-wide standards and tools for digitally signing and watermarking authentic media at the point of capture, enabling verifiable provenance and easy detection of alterations. Projects like the Content Authenticity Initiative (CAI) are leading this charge. * Advanced Detection AI: Continued investment in AI systems that can reliably detect synthetic media, moving beyond simple facial artifacts to more complex analyses of behavior, biometrics, and semantic consistency. * Privacy-Preserving AI: Research into AI models that can generate or analyze data without requiring access to sensitive raw personal information, thus minimizing the risk of misuse. 3. Platform Accountability and Moderation: * Rapid Takedown Policies: Mandating and enforcing stricter policies for platforms to expeditiously remove reported deepfakes, especially non-consensual explicit content. * Proactive Scanning: Implementing AI-driven systems to proactively scan for and identify potentially harmful deepfakes, rather than solely relying on user reports. * Transparency and Reporting: Improving transparency around content moderation decisions and providing clear, accessible channels for users to report and appeal content issues. 4. Public Awareness and Digital Literacy: * Educational Campaigns: Continuous public education campaigns to inform individuals about the dangers of deepfakes, how to identify them, and the importance of digital consent. * Critical Thinking Skills: Fostering critical thinking skills to empower individuals to question the veracity of online content and understand the potential for manipulation. 5. Ethical AI Development Practices: * Responsible AI Design: Encouraging AI developers to incorporate ethical considerations and safeguards against misuse from the very beginning of the development lifecycle. * "Red Teaming" and Vulnerability Testing: Proactively testing AI systems for potential malicious uses, including the creation of harmful deepfakes, and patching vulnerabilities before deployment. The protection of public figures and individual privacy in the AI era is not just about enacting laws; it requires a holistic ecosystem of technological innovation, legal enforcement, ethical development, and societal education. As AI capabilities expand, the imperative to safeguard digital identity and consent becomes ever more urgent.

As we navigate through 2025 and look further into the future, the intertwined destinies of AI, consent, and digital identity present both thrilling possibilities and formidable challenges. The "ai scarlett johansson sex" phenomenon serves as a grim precursor to a future where distinguishing between what is real and what is synthetically generated becomes increasingly difficult. This blurring of lines compels us to fundamentally rethink our relationship with digital content and, more importantly, with our own digital selves. The Evolving Nature of Digital Identity: Our digital identity is no longer merely a collection of profiles and data points; it is an increasingly dynamic and manipulable entity. AI’s capacity to replicate and distort our likeness, voice, and even mannerisms means that our digital identity can exist and act independently of our conscious control. This raises profound philosophical questions: who owns our digital likeness? What does it mean to have agency over our self-representation when AI can forge new realities? The concept of digital identity is evolving from a passive reflection to an active, often vulnerable, construct. The Expanding Scope of Consent in the Digital Realm: Traditional notions of consent, often rooted in physical interaction or explicit agreement for data sharing, are proving inadequate for the AI age. The mere existence of public images or voice recordings, even legitimately obtained, now presents a risk when AI can repurpose them for non-consensual creations. We must move towards a more expansive understanding of digital consent, one that explicitly addresses the use of an individual's likeness and data for synthetic media generation. This may involve: * Opt-in/Opt-out Mechanisms: Clear, granular controls for individuals to specify how their digital likeness and voice can be used by AI systems. * Legal Presumption of Non-Consent: A legal default that presumes non-consent for the creation of intimate or reputation-damaging deepfakes, placing the burden of proof on the creator. * Technological Consent Layers: Tools and platforms that allow individuals to embed digital consent markers into their own content, indicating permissible uses by AI. AI as a Tool for Empowerment vs. Exploitation: While the focus on deepfake misuse is critical, it's important to acknowledge AI's potential for empowering digital identity and expression. AI can enable new forms of artistic creation, personalized digital avatars, and immersive virtual experiences that enhance human connection and creativity. The challenge lies in fostering AI development that respects individual autonomy and privacy, channeling its power towards positive innovation rather than exploitation. This involves: * "Identity-Protective AI": Developing AI models specifically designed to protect digital identities, detect misuse, and empower individuals to control their synthetic representations. * Ethical Frameworks for Synthetic Media: Establishing global ethical guidelines for the creation and use of synthetic media, emphasizing transparency, attribution, and consent. * Educational Initiatives: Proactive education on the responsible use of AI and synthetic media, particularly for younger generations who are digital natives. The Imperative for Collaboration: The future of AI, consent, and digital identity cannot be shaped by any single entity. It requires continuous, robust collaboration among: * Governments: To enact comprehensive, enforceable laws that protect digital rights and hold malicious actors accountable. * Tech Companies: To develop ethical AI systems, build in robust safeguards, implement effective content moderation, and take responsibility for the societal impact of their technologies. * Civil Society Organizations: To advocate for victim rights, raise awareness, and provide support. * Academics and Researchers: To advance the science of AI ethics, detection, and identity protection. * The Public: To demand ethical AI, practice critical media literacy, and advocate for their digital rights. The trajectory of AI’s impact on consent and digital identity hinges on our collective ability to proactively address its risks while harnessing its potential. The choices we make now, in 2025, will determine whether AI becomes a tool that liberates and empowers or one that further erodes our privacy and autonomy in the digital age. The goal is to build a future where technological advancement and human dignity are not mutually exclusive, but rather mutually reinforcing.

The existence of "ai scarlett johansson sex" as a search term and the broader phenomenon of non-consensual deepfakes underscore a profound digital crisis that demands urgent and coordinated responsibility from all stakeholders. We are at a critical juncture where the lines between genuine and fabricated reality are rapidly dissolving, impacting individuals, institutions, and the very fabric of truth. Navigating this complexity requires more than just reactive measures; it calls for a proactive, ethical, and collaborative approach. Individual Responsibility: As individuals, we bear the responsibility to cultivate robust digital literacy. This means questioning the veracity of content we encounter online, especially highly emotional or sensational material. It means understanding the capabilities of AI to manipulate media and recognizing that seeing is no longer always believing. Furthermore, it involves practicing empathy and critical thinking before sharing content that could be fabricated or harmful. We must also be mindful of our own digital footprints, understanding how our publicly available data can be misused. For those who encounter deepfakes, the responsibility extends to reporting them to platforms and supporting victims, rather than contributing to their spread. It's about being informed digital citizens in 2025 and beyond. Platform Responsibility: Social media companies and content hosting platforms carry immense responsibility, given their role as gatekeepers of information and digital interaction. This responsibility includes: * Robust Content Moderation: Investing heavily in AI and human moderation teams capable of rapidly identifying and removing non-consensual synthetic media. * Transparency and Accountability: Being transparent about their content policies, enforcement mechanisms, and the volume of harmful content they address. * Proactive Measures: Moving beyond reactive takedowns to proactively develop and implement safeguards that prevent the creation and dissemination of deepfakes in the first place, perhaps through source authentication technologies. * Victim Support: Providing clear, accessible, and empathetic channels for victims to report abuse and seek redress, including efficient removal processes and potentially legal support resources. Developer and Research Community Responsibility: Those at the forefront of AI innovation have a moral and ethical obligation to consider the societal implications of their creations. This means: * "Ethics by Design": Integrating ethical considerations, including privacy, fairness, and accountability, into the very design and development phases of AI technologies. * Bias Mitigation: Actively working to identify and mitigate biases in datasets and algorithms that could inadvertently lead to harmful outcomes or disproportionately affect certain groups. * Responsible Disclosure: If developing powerful generative AI, working with policymakers and civil society to understand and manage potential misuse scenarios. * Investing in Detection: Directing research efforts not only into generating more realistic AI content but also into robust methods for detecting it and distinguishing it from genuine media. Government and Legislative Responsibility: Governments must act swiftly and decisively to create effective legal frameworks that protect individuals from the harms of deepfakes. This includes: * Clear Definitions: Crafting laws that precisely define non-consensual synthetic intimate imagery and criminalize its creation and distribution. * Enforceability: Ensuring these laws are enforceable across international borders and that law enforcement has the resources and expertise to investigate and prosecute perpetrators. * International Cooperation: Fostering global collaboration to address the cross-jurisdictional challenges of online harm. * Digital Rights: Championing digital rights that protect personal autonomy, privacy, and the right to one's own likeness in the age of AI. The "ai scarlett johansson sex" issue is not an isolated incident but a symptom of a larger societal challenge concerning the responsible development and deployment of AI. Our collective response to this challenge will define the future of trust, truth, and personal sovereignty in the digital realm. It is a shared responsibility, demanding continuous vigilance, adaptability, and an unwavering commitment to ethical principles. By embracing this call for responsibility, we can strive to build a digital future where innovation thrives without compromising human dignity and safety. url: ai-scarlett-johansson-sex keywords: ai scarlett johansson sex