#70 Analysing the Global Response to Deepfakes
Fake It Till You Legislate It, Why are the Artemis Accords Controversial?
Today, Rohan Pai discusses the recent flurry of policy responses to deepfakes around the world.
Aditya Ramanathan highlights the reason why the Artemis Accords are controversial with an excerpt from a new Takshashila Discussion Document.
Also,
We are hiring! If you are passionate about working on emerging areas of contention at the intersection of technology and international relations, check out the Staff Research Analyst position with Takshashila’s High-Tech Geopolitics programme here. For internship applications, reach out to satya@takshashila.org.in.
Cyberpolitik: Fake It Till You Legislate It
— Rohan Pai
On the morning after Christmas 2023, Instagram, Facebook, X, YouTube, and Google, amongst other social media giants, woke up to a stern advisory by the Union Government that cautioned them against hosting content created using deepfake technology.
This advisory, in a nutshell, orders that these digital platforms or ‘intermediaries’ be explicit in their communication with users about their zero tolerance for certain types of content, including deepfakes, listed under Rule 3(1)(b) of the IT Rules published by MeitY. Such an advisory effectively shifts the responsibility of deepfake legislation from the sarkaar onto the bazaar. Rather than the actual perpetrators of deepfakes, it is now the intermediaries who face the risk of suspension, thereby losing access to the lucrative Indian market of a potential billion users.
Deepfake legislation, however, is by no means an Indian problem. The dystopian reality of misinformation transmitted via AI-generated impersonations, the likeness of which only continues to grow exponentially, has left world leaders everywhere clutching at straws. The pioneer that emerged with what was possibly the first-ever comprehensive and nationwide policy on deepfake legislation was none other than the People’s Republic of China. While often dubbed as ‘The Great Firewall’, even China’s robust internet censorship mechanisms have struggled to monitor deepfakes, thus prompting firm legislation.
Released in late 2022 by the CAC, the regulations not only mandate that consent from individuals be sought prior to using their face in a deepfake but also that watermarks be used as a way of filtering deepfakes from authentic videos. Notably, the regulations do not outrightly ban the usage of deepfake technology, but the clause requiring consent means that the Chinese are unlikely to see Xi Jinping deepfakes on their Weibo feeds anytime soon.
Moving across the Pacific to the land of the free, censorship is replaced by its extreme liberal counterpart: The First Amendment. Ever since its enactment in 1791, the First Amendment’s protection of free speech has captured the American imagination and continues to be something of a cultural touchstone in U.S. policymaking. Against such a backdrop, any attempts at policing the exchange of thoughts and ideas, even if born out of noble intentions, are quick to ruffle feathers.
Of course, concerns surrounding deepfakes are very much prevalent in the U.S., but these crop up primarily during the election season. Just last week, Joe Biden’s voice was manipulated in a deepfake ‘robocall’ urging Democrats not to vote in the New Hampshire primaries. Unlike China, however, where anti-regime sentiments form the basis of deepfake fears, party politics in the U.S. has long been characterised by satirical commentary and ad hominem attacks between candidates. Ron DeSantis’ campaign video, for example, that used AI-generated images of Donald Trump kissing and embracing Dr. Anthony Fauci, serves as a testament to deepfake technology fitting neatly into the current political climate.
It is, thus, unsurprising to find that deepfake regulations in the U.S. are restricted to a select few states, and even these are relatively mild. Considering that the internet is accessible nationwide, disparate state laws seem to prove ineffective in curbing the spread of misinformation.
Having analysed both ends of the spectrum, India’s stance on deepfake legislation appears ideologically closer to that of China. In fact, Union Minister Rajeev Chandrasekhar warned that the regulations in the new advisory would become amended into national law should intermediaries fail to comply with them.
While a hardline approach may promise greater efficacy, as demonstrated by the recent arrest of a celebrity deepfake creator in Delhi, it poses several threats that are not readily apparent. Corrupt politicians, for instance, if found to be guilty via incriminating video clips, may reasonably allege them to be deepfakes as a means of escaping public censure.
Stricter rules also risk pigeonholing deepfake technology as an inherently malicious usage of artificial intelligence, which is far from the truth. Whether by breathing life into historical figures, enabling influential speakers to communicate in multiple languages, or even providing anonymity to journalists reporting from war-torn regions, deepfakes boast a multitude of benefits in the arena of social welfare.
Fear has always been known to accompany the arrival of new technology, and deepfakes are simply the latest double-edged sword caught in the crossfire.
India is set to host the Quad Leaders' Summit in 2024. Subscribe to Takshashila's Quad Bulletin, a fortnightly newsletter that tracks the Quad's activities through the Indo-Pacific.
Your weekly dose of All Things China, with an upcoming particular focus on Chinese discourses on defence, foreign policy, tech, and India, awaits you in the Eye on China newsletter!
The Takshashila Geospatial Bulletin is a monthly dispatch of Geospatial insights for India’s strategic affairs. Subscribe now!
Antariksh Matters: Why are the Artemis Accords Controversial?
— Aditya Ramanathan
We’ve published a new discussion document that provides an overview of the existing international and domestic provisions for governing human activity on the Moon. The international treaties go back all the way to the 1963 Partial Test Ban Treaty (PTBT), which prohibited nuclear testing in several environments, including celestial bodies. The domestic provisions have more recent provenance, being less than a decade old.
The most controversial recent international initiative is the Artemis Accords, a set of non-binding guidelines for conduct on celestial bodies. The accords are primarily meant to provide a series of guidelines for the US-led Artemis lunar exploration programme. They were launched in October 2020 by eight states: the US, the United Kingdom, Australia, Canada, Italy, Japan, Luxembourg, and the UAE. As of January 2024, there were a total of 34 signatories, including India. Below is a lightly edited excerpt from the document:
The Accords
The Artemis Accords lay out ten principles for activity on celestial bodies. Six of these essentially restate core provisions of the 1967 Outer Space Treaty (OST): cooperating for peaceful purposes, transparency, interoperability, emergency assistance, registration of space objects, and the release of scientific data.
Another provision calls for managing orbital debris and ensuring the safe disposal of spacecraft. While this is not explicitly codified in international law, debris mitigation is part of the practices being consciously adopted by most spacefaring states.
However, it is the three remaining provisions of the accords that have been the source of controversy. These are: protecting heritage in space, deconflicting activities, and allowing the extraction and use of resources.
The Controversial Provisions
Protecting Heritage: The provision for protecting heritage is primarily intended to preserve the historic Apollo landing sites from harm. However, this seemingly innocuous provision could lead to de facto claims of national territory. While the OST clearly prohibits claims of sovereignty in space, the accords provide no clear definitions of what constitutes heritage and prescribe no time limits.
The Artemis Accords simply defines outer space heritage “to comprise historically significant human or robotic landing sites, artefacts, spacecraft, and other evidence of activity on celestial bodies”. This definition would, for instance, cover the landing site of India’s Chandrayaan-3 mission. However, it should be noted that the accords clearly indicate that such declarations should not be unilateral and must be carried out in “accordance with mutually developed standards and practices.” Artemis signatories are also to use their experience “to contribute to multilateral efforts to further develop international practices and rules”.
Deconflicting: This provision cites Article IX of the OST, which covers “harmful interference,” a principle that remains unclearly defined. Artemis signatories are to avoid acts that constitute harmful interference with the activities of other signatories. If one signatory believes that the activities of another signatory constitute harmful interference, they are “to provide each other with necessary information regarding the location and nature of space-based activities”.
To avoid harmful interference, signatories can declare safety zones based on four principles for safety zones:
· The size and scope of the safety zone must be based on the activity being undertaken.
· The size and scope must be “determined in a reasonable manner leveraging commonly accepted scientific and engineering principles”.
· Signatories must change the size and scope of the safety zone as their activities change. It adds that safety zones “will ultimately be temporary, ending when the relevant operation ceases”.
· Signatories are to keep the UN Secretary-General informed of “establishment, alteration, or end of any safety zone, consistent with Article XI of the Outer Space Treaty”.
The deconfliction provision emphasises its adherence to the OST, arguing that its primary purpose is to avoid harmful interference. The principle also emphasises the temporary nature of safety zones, without prescribing actual time limits. Therefore, a safety zone declared around a Moon base may last indefinitely, for decades or even longer.
Space Resources: For something that has attracted much controversy, the provision on resource extraction is brief, making only four points:
· Space resource utilisation “can benefit humankind by providing critical support for safe and sustainable operations”.
· Extraction and utilisation must comply with the OST and must be “in support of safe and sustainable space activities”.
· In accordance with the OST, signatories must keep the UN Secretary-General informed of these activities.
· Signatories will use the experience gained to contribute to multilateral efforts on “international practices and rules” in this regard.
The most significant aspect of this provision is the statement that resource utilisation must be “in support of safe and sustainable space activities”. This statement is naturally open to interpretation, in particular since a wide range of undertakings can fall under the rubric of “space activities”. If “space activities” means supporting and sustaining human presence on the Moon or beyond, this provision strongly suggests in-situ resource utilisation (ISRU) rather than commercial mining of lunar minerals.
Lunar regolith could be used to build human habitation on the Moon, either by processing it into construction material or by simply using regolith to cover up human-made structures and thus provide an additional layer of protection from radiation and temperature variations. More crucial than regolith is water. If lunar water ice can be accessed and purified in-situ, it could be used to sustain humans as well as provide hydrogen and oxygen to power rockets. In theory, other lunar resources including metal ores could be mined for use on the Moon. However, such applications are unlikely to be feasible before mid-century.
The broad term “space activities” can also cover commercial mining for return-to-Earth (RTE) applications. The challenges with mining on celestial bodies for RTE are not just technological, but also economic. Space transport costs alone dwarf the inherent value that minerals like platinum, gold, or rare earths hold.
However, lunar mining for RTE can make sense if what is being mined barely exists on Earth. One potential example is Helium-3, which could have potential in future nuclear fusion technology.[ix] However, even if they become feasible, any such applications are likely to take decades to become reality.
While the Artemis Accords are not formal law, they build on the foundation of the OST to provide American lunar explorers the greatest possible freedom of action. As such, they are likely to cause varying degrees of misgivings for other major spacefaring states.
What We're Reading (or Listening to)
[Podcast: All Things Policy] AI in Mental Healthcare ft. Srishti Srivastava (CEO and Founder, Infiheal)
[Opinion (Times of India Blog)] Republic Day reminder: Let’s reclaim the right to economic freedom, by Anupam Manur and Pranay Kotasthane
[Opinion] The need to overhaul a semiconductor scheme, by Satya S Sahu and Pranay Kotasthane