AI Transformation in Bangladesh: Deepfake Tech and Artificial Intelligence Policy 2024
Navigating the Legal Terrain in the Age of AI: Challenges, Regulations, and the Bangladeshi Perspective
The recent surge in AI’s popularity has made household names out of tools like ChatGPT, Midjourney, and Quillbot. These AI marvels can do almost anything, from making Cristiano Ronaldo sing Hindi songs to merging faces of MKBHD and Shakib Al Hasan seamlessly. With AI at our fingertips, even the most fantastical scenarios come to life. Picture Cyberpunk-themed Dhaka cityscapes conjured by Midjourney, existing solely in our imaginations but made real through AI. As AI continues to evolve, it constantly challenges the limits of creativity and innovation. In this new era, the only boundary is our imagination.
However, currently the legal landscape finds itself at a crossroads, grappling with unprecedented challenges while striving to adapt to the rapid advancements in technology. From deepfake technology to algorithmic biases, the legal fraternity faces multifaceted dilemmas that demand innovative solutions and robust regulatory frameworks. This article explores the implications of AI on the legal sector, delves into the measures adopted by first-world countries, and sheds light on South Asia and Bangladesh’s initiatives in this realm.
The Deepfake Predicament
Deepfake technology, fueled by AI algorithms, poses a significant threat to the integrity of legal proceedings and personal privacy by seamlessly superimposing individuals’ faces onto fabricated videos or audio recordings. This not only undermines the credibility of evidence but also raises concerns about the manipulation of public opinion, scams and the erosion of trust in the judicial system. Beyond its implications for evidence tampering and misinformation, deepfake technology has been exploited for creating non-consensual pornography, commonly known as “deepfake porn.” This insidious practice involves superimposing the faces of individuals onto explicit content, often without their consent, leading to severe psychological harm, reputational damage, and violation of privacy rights. A report by Sensity AI, The State of Deepfakes 2019 Landscape, Threats, and Impact, found that 96 percent of deepfakes were non-consensual sexual deepfakes, and of those, 99 percent were made of women. (Dunn, 2021) We’ll use one of the top deepfake porn sites as an example. In 2018, the site contained less than 2,000 deepfake pornographic videos. This number grew to 13,000 by 2022, and the site is now attracting 16 million viewers per month. (Mft, 2023).
This issue does not stop here, as the sophistication of AI-generated deepfake voices further complicates matters, enabling scammers to perpetrate elaborate schemes with alarming ease. The recent case reported by The Wall Street Journal highlights the alarming reality: a CEO of a UK-based energy firm fell victim to an elaborate scheme where he believed he was conversing with the chief executive of the firm’s German parent company. In reality, the voice on the other end of the line was a fraudster utilizing AI voice technology to spoof the German CEO. Despite the CEO recognizing subtle cues in the fake voice, he followed orders to transfer €220,000 (approx. $243,000) to a bank account belonging to a Hungarian supplier. (Damiani, 2024).
Another manifestation of this issue is the proliferation of fake AI advertisements. YouTube sensation MrBeast, also known as Jimmy Donaldson, recently raised concerns about the proliferation of AI deepfake scams after a fake TikTok ad surfaced featuring his likeness offering $2 iPhones. Donaldson’s call for action highlights the growing threat posed by deepfake technology, which replaces a person’s face with someone else’s, often with alarming realism. While the ad has since been removed, its existence underscores the need for social media platforms to address the spread of deceptive AI-generated content. Donaldson’s experience echoes similar sentiments expressed by other public figures, including actor Tom Hanks and Zelda Williams, who have also spoken out against unauthorized AI recreations of themselves (MrBeast Calls TikTok Ad Showing an AI Version of Him a “scam,” 2023). Similarly, a manipulated video featuring former captain of the Bangladesh national cricket team, Shakib Al Hasan, surfaced on Facebook, falsely promoting a betting site. In the video, Shakib is shown endorsing the opening of a government-licensed mobile casino in Bangladesh, enticing viewers with significant bonuses. However, upon verification, it was revealed that the video had been edited, splicing together fragments from an original promotional video for Nagad, a digital financial service. The doctored reel was then circulated widely, garnering reactions and shares on social media. Notably, this isn’t an isolated incident; similar fake videos have targeted other Bangladeshi cricketers like Mustafizur Rahman. Such deceptive practices underscore the need for vigilance against fraudulent campaigns and the responsible use of digital media platforms. (Shakib Al Hasan’s Fake Video Used in Promoting Betting Sites on Facebook, 2024).
Share this Article
The Legal Reaction
In response to these challenges, international bodies and some countries are developing regulatory frameworks to mitigate AI-related risks.
The European Union, for instance, is pioneering a risk-based approach to AI regulation, acknowledging the imperative of adapting legal structures to the AI era. The European Union’s parliament has approved the world’s first major regulatory framework for artificial intelligence (AI), dividing AI technologies by risk level and implementing regulations from “unacceptable” to low hazard. This move positions Europe as a global standard-setter in AI governance, aiming to balance innovation with fundamental rights. (Gilchrist & Iordache, 2024). Stanford University’s AI Index for 2023 revealed that in 2022, 37 bills related to AI were enacted into law worldwide, with the United States taking the lead by passing nine of these laws. (How AI Is Transforming the Legal Sector, 2024). A California bill introduced on 8th February would require licensing of powerful AI models, via a new state agency and would create an AI public computing cluster. (Bill Text – SB-1047 Safe and Secure Innovation for Frontier Artificial Intelligence Models Act., n.d.) South Dakota Republican Gov. Kristi Noem has signed an updated law against child sexual abuse images on 12th February, mandating prison sentences for those caught creating, distributing and possessing AI-generated images. (Gov. Noem Signs Bills Into Law – News, n.d.).
The National Perspective
While strides are being made on a global scale, attention must also be directed to South Asian countries like Bangladesh, where the rapid escalation of AI-powered crimes necessitates proactive measures. Bangladesh has recently crafted the National Artificial Intelligence Policy 2024 (AI Policy 2024) to address the multifaceted challenges of AI, spanning social, legal, and ethical domains. Key features include the establishment of the National Artificial Intelligence Center for Excellence (NAICE) and a proposed National AI Advisory Council to coordinate efforts across government, academia, industry, and civil society. While the policy prioritizes sectors like education and healthcare, it lacks specific provisions for defense and national security. Emphasizing the need for robust international collaboration, Bangladesh aims to combat emerging risks such as organized cybercrimes. Drawing insights from India’s strategy, Bangladesh underscores the importance of partnerships for AI research and innovation. However, achieving a balance between AI integration and risk mitigation poses practical challenges, necessitating investments in infrastructure, compatible laws, and collaboration between academia and industry. (Baree, 2024).
Conclusion
The intersection of AI and the legal sector presents both opportunities and challenges, from the threat of deepfake technology to the need for robust regulatory frameworks. As international bodies and first-world countries forge ahead with innovative approaches to AI governance, South Asian nations like Bangladesh are also taking proactive steps to address emerging risks. Through initiatives like the National Artificial Intelligence Policy 2024, Bangladesh aims to navigate the complexities of AI integration while safeguarding against potential harms. As we continue to chart this new frontier, collaboration and vigilance will be crucial in shaping a future where AI serves as a force for good in the legal landscape.
References:
- Dunn, S. (2021, March 3). Women, not politicians, are targeted most often by deepfake videos. Centre for International Governance Innovation. https://www.cigionline.org/articles/women-not-politicians-are-targeted-most-often-deepfake-videos/
- Mft, J. N. M. (2023, November 16). The damage caused by deepfake porn. Healthnews. https://healthnews.com/mental-health/anxiety-depression/the-damage-caused-by-deepfake-porn/
- Damiani, J. (2024, February 20). A voice deepfake was used to scam a CEO out of $243,000. Forbes.
https://www.forbes.com/sites/jessedamiani/2019/09/03/a-voice-deepfake-was-used-to-scam-a-ceo-out-of-243000/?sh=5024de732241 - MrBeast calls TikTok ad showing an AI version of him a “scam.” (2023, October 3). NBC News.
https://www.nbcnews.com/tech/mrbeast-ai-tiktok-ad-deepfake-rcna118596 - Shakib Al Hasan’s fake video used in promoting betting sites on Facebook. (2024, February 28). Dismislab.
https://en.dismislab.com/shakib-al-hasans-fake-video-betting-sites/ - Gilchrist, K., & Iordache, R. (2024, March 13). World’s first major act to regulate AI passed by European lawmakers. CNBC. https://www.cnbc.com/2024/03/13/european-lawmakers-endorse-worlds-first-major-act-to-regulate-ai.html
- How AI is transforming the legal sector. (2024, February 22). RSM Global. https://www.rsm.global/insights/how-ai-transforming-legal-sector
- Bill Text – SB-1047 Safe and Secure Innovation for Frontier Artificial Intelligence Models Act. (n.d.). https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202320240SB1047
- Gov. Noem Signs Bills into Law – News. (n.d.). https://news.sd.gov/news?id=news_kb_article_view&sys_id=2db36c4a8710c210428042e8cebb35f9
- Baree, M. N. M. (2024, April 18). An overview of Bangladesh National Artificial Intelligence Policy 2024. The Daily Star. https://www.thedailystar.net/law-our-rights/law-vision/news/overview-bangladesh-national-artificial-intelligence-policy-2024-3590351