News Briefing 18-22 September

Autonomy in weapons systems, and AI in security and defence

• International: World needs ‘Statesmanship, Not Gamesmanship and Gridlock’, Secretary-General Tells General Assembly, Calling for Determination to Uphold Charter’s Pledge for Peace: In his opening address at the 78th United Nations General Assembly, UN Secretary General Antonio Guterres told the General Assembly that ‘Determination for peace also requires new governance frameworks for emerging threats, from artificial intelligence to lethal autonomous weapons [systems] that function without human control.’
• International/Switzerland: AI racing drone could transform warfare – but not yet, says maker: Forbes carries a report on the work of Davide Scaramuzza and his colleagues at the University of Zurich’s Robotics and Perception Group, whose AI-enabled racing drones have the potential to ‘transform’ warfare. Scaramuzza himself notes that the results of his tests with racing drones ‘should open our eyes on what AIs can potentially enable if unleashed’ and that ‘that’s why in numerous talks at the United Nations, I always argue that we need a Geneva convention to stop killer robots.’
• Europe/Israel: Elbit sells SkyStriker loitering munition to mystery European customer: The Israeli company Elbit Systems ‘will supply several hundred canister configuration SkyStriker loitering munitions to an undisclosed European country in a $95 million deal’, reports Defense News. Elbit describes the SkyStriker as ‘a fully autonomous loitering munition (LM) that can locate, acquire and strike operator designated targets with a 5 or 10 Kg warhead installed inside the fuselage, enabling high-precision performance.’ Elbit also states that the system ‘can precisely strike targets while maintaining a “man in the loop” even in GPS and communication denied environments.’
• US: DARPA is funding AI to help make battlefield decisions: Live Science reports on the U.S. Defense Advanced Research Projects Agency (DARPA) project SCEPTER (Strategic Chaos Engine for Planning, Tactics, Experimentation and Resiliency), with DARPA hoping that ‘military simulations will become less computationally intensive, which, in turn, could lead to better, quicker war strategies.’ Meanwhile, Breaking Defense highlights a recent report arguing that the Department of Defense should explore the possibility of having large language models (LLMs) create military operations order.
• US: A.I. and the next generation of drone warfare: A New Yorker report on the U.S. Department of Defense’s recently announced Replicator programme notes that whether the new AI-assisted and autonomous technologies envisioned in the Replicator programme ‘remain restricted to the battlefield, or get repurposed in new and unbridled ways—for example, in domestic surveillance by an increasingly militarized police force—remains to be seen.’
• US/South Korea: U.S. Marines evaluate South Korea’s new robotic combat vehicle: South Korean defence conglomerate Hanwha has ‘signed a contract with the U.S. Department of Defense for the Foreign Comparative Performance Test (FCT) project’, and will test the Arion-SMET, an unmanned ground vehicle, at the Marine Corps Training Center on the island of O’ahu, Hawaii, for three weeks from early December this year.
• US: New Ghost UAS variant increases payload capacity & flight time: Anduril Industries has introduced the newest variant of its Ghost UAS, with a longer flight time and additional payload options. Anduril Ghost systems allegedly possess ‘onboard compute with advanced computer vision algorithms to autonomously detect, classify, and track objects of interest while intelligently navigating the terrain and airspace’.
• Germany/ Ukraine: Ukraine Promised Drone Boat Fleet By Germany. It May Be Weapons Not Yet Seen: Forbes, in this piece, highlights the proven utility of USV in the Russia- Ukraine battle. It reports on the announcement by German Parliamentary State Secretary Siemtje Möller promising ‘50 surface drones are designed to help throw back the aggressor not only on land, but also at sea’ to Ukraine.’ The report further mentions that Germany ‘may be planning to produce a completely new range of small USVs from a clean slate.’ 

Facial recognition, biometric identification, surveillance 

• Argentina: The Twisted Eye in the Sky Over Buenos Aires: WIRED reports on the widespread use of facial recognition technologies by law enforcement agencies in Argentina. This system has led to around 140 cases of wrongful arrest and further judicial enquiry uncovered that ‘15,459 persons were loaded into the facial recognition system without a request from the judiciary, i.e. without a legal basis for doing so.’ Despite three separate judicial pronouncements declaring these systems illegal, the law enforcement agencies have been making efforts to restart its use.
• International: How a ‘Digital Peeping Tom’ Unmasked Porn Actors: This piece by WIRED covers the risks of publicly available facial recognition algorithms that can track a person through social media, and a myriad other forms of digital presence. The impact of such algorithms for professions such as porn workers causes unprecedented harm and invades privacy of the individual.

AI, algorithms and autonomy in the wider world

• International: The AI heretic: How a leading economist learned to start worrying and fear artificial intelligence: Business Insider carries an interview with Daron Acemoglu, an economist at MIT, who notes that ‘humans have greatly benefited from technology, but there’s nothing automatic about that… It could have gone in a very bad direction had it not been for institutional, regulatory, and technological adjustments. That’s why this is a momentous period: because there are similar choices that need to be made today.’
• Finland: These Prisoners Are Training AI: Wired reports on the use of data labor from Finnish prisons  to train artificial intelligence for private companies. While some claim that data labor in prisons prepares people inside for the digital age workforce once they are released, Amos Toh, a senior researcher focusing on artificial intelligence at Human Rights Watch says that ‘labor behind building tech is being outsourced to workers that toil in potentially exploitative working conditions.’ 
• International: Data Work and its Layers of (In)visibility: This pertinent piece by Just Tech focuses on the emerging concept of ‘Ghost work’ where labor that goes into building technology is unrecognized and invisibilized.  It covers the ‘the unpaid work performed by users, human workers pretending to be AI systems, and the different forms of exploitation of vulnerable communities globally.’ It calls for actions to rectify and repair in addition to exposing injustice through transparency in how ‘data is utilized, the opportunity to seek fair compensation for our work, and the ability to opt in or out of assisting in AI development’ 
• International: Is it possible to regulate artificial intelligence?: BBC reports on the need for AI regulation and the utility of pursuing an international regulation through the UN. While there are numerous efforts to regulate AI at the UN level, Jimmy Wales, the founder of Wikipedia suggests that this would be difficult to achieve. 
• Australia: Australian federal police using AI to analyse data obtained under surveillance warrant: The Guardian reports that the ‘Australian federal police has said it uses AI to analyse data obtained under telecommunications and surveillance warrants.’ It further says that though the use of AI has been limited to translation of foreign materials into English, it provides ‘opportunity to find useful information in large, lawfully collected datasets.’ 

Research and reports

• Study of image sensors for enhanced face recognition at a distance in the Smart City context: Biometric Update reports on a research paper published in Nature on a method to ‘optimize surveillance cameras for matching face biometrics from a distance’ based on deep learning of data including ‘distance of the individuals from the camera, the focal length of the image sensors and the size in pixels of the target face.’  The method claims an accuracy of 99 percent, and is believed to be ‘crucial in security applications in smart cities.’