Artificial intelligence ("AI") has become ubiquitous, perhaps even workaday in certain spheres and it is undeniable that it is transforming a few sectors. Through inter alia algorithms and deep learning AI is becoming interconnected, predictive, adaptive, emergent and autonomous. It is fascinating that responsibility, ethics and transparency within the AI ecosystem are being enhanced without compromising innovation. Whilst AI software, applications and technologies become increasingly integrated into our daily lives, a myriad of issues like privacy, copyright, data security, sensitive information and ethical use have been accentuated.
Ghibli art - recent furore and fervour
The recent fervid use of Ghibli art poses critical questions about infringement of copyright and moral rights, as artists/creators seek to protect their original works from unauthorized replication or even disambiguation or defacement. For instance, OpenAI's DALL-E 3 feature allows users to create images in the distinctive art style of Hayao Miyazaki1 or Studio Ghibli. Thousands of AI-generated images imitating the visual style of Studio Ghibli flooded social media and this trend has been driven by tools on ChatGPT and Grok, which can transform any image into dreamlike scenes with soft palettes and ethereal landscapes, characteristic of classic works. For Miyazaki (Studio Ghibli's founder), art should come from the spirit of the author, not from algorithms.
Though an artist's style as such is not protected by intellectual property law, the works that represent it are covered. According to the Berne Convention2, the reproduction/adaptation/transformation of a work requires the authorisation of its creator or owner. If AI models were trained on Ghibli's copyrighted images without consent, it would make a prima facie case of infringement at least and even a defence of 'fair use' could prove to be demonstrably fallacious. Inevitable questions arise: can an AI-generated image constitute a derivative work; should companies developing AIs obtain licences; should they be paying royalties? Perhaps a line needs to be drawn between legitimate inspiration and infringement.
Clearview AI controversy
The Clearview AI controversy puts the spotlight on ethical challenges of facial recognition technology, inviting regulatory scrutiny in light of potential privacy violations. Clearview AI, a US-based facial recognition company, has sparked global controversy for its practice of harvesting billions of publicly available images from social media, websites, often sans consent, to generate a massive biometric database. Its technology, retailed essentially to law enforcement agencies (accessed by more than 600 entities in US, Canada, including during the Russia-Ukraine conflict), cross-references uploaded photos against a database of over 10 billion faces scraped from platforms like Facebook, Twitter, Google, and VKontakte.
It is an increasingly popular opinion that Clearview's methods violate individual privacy rights, as images of individuals, including those deleted from public domains, are permanently stored and analysed without their knowledge, creating irreversible risks of surveillance overreach.3 Advocates for civil rights and privacy also highlighted escalating risks of misidentification as facial recognition tools and software demonstrate racial and/or gender bias, leading to harassment of civilians and sometimes even wrongful arrests. Countries like Australia and the European Union took stringent steps as they commenced investigations or implemented bans against Clearview, while Facebook and X (erstwhile Twitter) also issued cease and desist orders to Clearview for violation of their data policies.
Apple's AI - doctor in your pocket
Novel AI related challenges are becoming more and more apparent in sectors such as healthcare as well. Apple's AI initiatives, often referred to as the "doctor in your pocket" collect sensitive health related data, necessitating compliance with stringent regulations such as the HIPAA4. Apple's healthcare AI initiatives, which integrate features like Health Records, predictive health analytics, and wellness tools into the Health app and Apple Watch, have raised serious concerns. Even as Apple continues to reiterate safeguards (like end-to-end encryption) on devices and iCloud, a plethora of concerns persist such as ethical AI governance. Shortfalls and imperfections in privacy, data protection, and potential IP violations have been persistent concerns.
In spite of Apple's mandatory requirement for privacy policies, exposure of medical information or personal sensitive data through third party applications and software is a glaring and perilous risk. Add to that, specialized features such as Medical ID which grant emergency response teams access to critical health data from locked devices, and the ethical implications of health predictions made using generative AI applications.5 Few can deny the expeditious resolutions (like on the spot diagnosis in an emergency) and the promise of the greater common good that such advancements aim at, but fewer still can refute the legal ramifications, ethical conundrums and complicated debates that inevitably accompany these endeavors with AI at the core.
Corollary
The concerns surrounding AI today are more nuanced than just privacy/sensitive data issues (like the 23andMe data breach6 in October 2023, which led to a class action lawsuit, with the company eventually agreeing to pay $30 million in damages). In view of the illustrations above, risks ahead demand avoidance of equivocation in legislations, rules of procedure and judicial precedents. The legal landscape is further complicated by multiple regulatory approaches across jurisdictions.
The CCPA7 (in the USA) has set a precedent for consumer rights, while the GDPR8 implemented in EU establishes stringent standards for data protection. India's DPDP Act9 aims to create a similar framework, highlighting the worldwide trend towards enhanced data governance. Hardly any technology expert, AI developer, academician or legal luminary can look past the apparent benefits of AI to economies, industries and to civil societies. However, as AI becomes more sentient and sophisticated, we cannot remain obtuse to the emergent challenges.
Footnotes
1 Renowned Japanese animator and filmmaker
2 Berne Convention for the Protection of Literary and Artistic Works 1886
3 https://www.business-humanrights.org/en/latest-news/clearview-ais-facial-recognition-technology-designed-for-surveillance-of-marginalized-groups-report-reveals/
4 Health Insurance Portability and Accountability Act of 1996 (US Federal law)
5 About the privacy and security of your health records - Apple Support (2024),https://support.apple.com/en-us/111755
6 a prominent provider of genetic testing services
7 California Consumer Privacy Act 2018
8 General Data Protection Regulation - Regulation (EU) 2016/679 effective from 25 May 2018
9 Indian Digital Personal Data Protection Act, 2023 (Procedural Rules to be notified).
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
