News Briefing 1-5 April 2024

Autonomy in weapons systems, and AI in security and defence

• Gaza/Israel: ‘Lavender’: The AI machine directing Israel’s bombing spree in Gaza: +972 Magazine and Local Call reports on Israel’s ‘Lavender’ system, a data processing system which, according to six Israeli intelligence officers, ‘has played a central role in the unprecedented bombing of Palestinians, especially during the early stages of the war.’ The Lavender system generates human targets, and +972 reports that ‘human personnel often served only as a “rubber stamp” for the machine’s decisions, adding that, normally, they would personally devote only about “20 seconds” to each target before authorizing a bombing — just to make sure the Lavender-marked target is male.’ Read Stop Killer Robot’s comment on Lavender and AI target recommendation systems here. Al Jazeera carries comment on the Lavender system from Automated Decision Research researcher Sai Bourothu here. The Lavender system is also covered here in The Guardian, and here in The Washington Post.  
• Israel: Israeli startups hope to export battle-tested AI tech: Nikkei Asia reports that ‘Israeli startups are looking to export advanced military technology powered by artificial intelligence’ whose precision has allegedly  ‘improved from use in the Israel-Hamas war.’ The piece quotes Axon Vision CEO Roy Riftin, who says that his company (which manufactures Edge360, an AI system which is installed in Israeli armoured vehicles currently in use in Gaza) has a ‘close relationship’ with the Israeli army.
• US/Australia: Shield AI moves into Australia with Sentient Vision Systems deal: Shield AI (a US based company) has announced that it has entered into an agreement with Sentient, an Australian based company. The move is intended to merge the expertise of Shield AI’s ‘artificial intelligence pilot, drones and Sentient’s imagery capabilities.’ The two companies have been working together since last year.
• US: Robotic moves: Army picks 8 tech companies to compete Robotic Combat Vehicle pieces: The Defense Innovation Unit (DIU) has announced that a number of companies have been identified to develop ‘robust, capable, and compliant software systems that can operate in a variety of autonomous modes and rapidly integrate a variety of payloads.’
• International: Thales Launches Defense-Focused AI Platform: Find Biometrics reports on the launch of a new AI platform by defence company Thales, named CortAIx. The new platform features systems for faster live processing of data from sensors as well as improved surveillance and swarming capabilities of existing systems. 

AI, algorithms and autonomy in the wider world

• International: Billie Eilish, Nicki Minaj, Stevie Wonder and more musicians demand protection against AI: Hundreds of musicians around the world have signed a letter stating their concerns around the invasion of artists’ privacy as well as ‘the use of their identities without their consent, and the use of their music to train AI models.’They also mentioned that ‘These efforts are directly aimed at replacing the work of human artists with massive quantities of AI-created ‘sounds’ and ‘images’ that substantially dilute the royalty pools that are paid out to artists.
• South Africa: South Africa requires comprehensive AI legislation – experts: ‘South Africa currently has no explicit legal framework for AI despite showing an eagerness to be a key player in the space’.  ICT legal experts at ENS Africa raise the issue of lack of transparency within the sector that can certainly lead to ‘potential biases, incorrect decisions, and inappropriate models.’Having outlined the risks, experts have firmly expressed that it is ‘critical to implement comprehensive legislation to protect the best interests of society.’
• International: ‘Many-shot jailbreak’: lab reveals how AI safety features can be easily bypassed: The Guardian reports on a new research by AI lab Anthropic on generative AI platforms. It uses a method known as ‘jailbreak’ to trick safety measures in generative AI platforms which are designed to ‘encourage it to refuse certain requests’ which may be illegal or harmful.

Facial recognition, biometric identification, surveillance 

• International: The changing face of protest: Rest of World has published a new report on the use of facial recognition technologies to surveil protests around the world, underscoring that ‘while authorities generally pitch facial recognition as a tool to capture terrorists or wanted murderers, the technology has also emerged as a critical instrument in a very particular context: punishing protesters.’ 
• US: FTC Rejects ESRB’s Proposal to Use Facial Recognition Age Verification Tool: Last year, the Entertainment Software Ratings Board (ESRB) requested approval from the Federal Trade Commission (FTC) for the use of “Privacy-Protective Facial Age Estimation’. The ESRB argues that the tool has adequate guardrails but the request has been denied by the FTC and the ESRB could possibly re-submit a request in the near future. 
• US: White House AI policy a ‘seismic policy shift’ for biometrics developers and vendors: Biometric Update reports on World Privacy Forum’s analysis of the White House’s new AI policy, lauding it as the ‘most assertive, clear, and direct articulation of biometric governance by the U.S. government to date.’ The analysis makes specific mention of the provisions in the policy requiring disclosure of consent for collection of biometric data.
• Russia: Leaked documents reveal details on Russia’s upcoming surveillance system: This piece in Biometric Update reports on new documents uncovering Russia’s investment into surveillance systems. As part of this project, surveillance cameras for crowd management, licence recognition software and facial recognition software which are designed to ‘build biometric templates of people’s faces’ will be deployed nationally.  

Research and reports

• Opinio Juris symposium on military AI and the law of armed conflict: The academic international law blog Opinio Juris features a symposium on military AI and the law of armed conflict, with a range of pieces on the subject, including a piece from the ICRC on artificial intelligence in military decision making, noting that AI decision support systems (such as the Lavender system) ‘must only be applied within legal, policy and doctrinal frameworks that ensure respect for international humanitarian law. Otherwise, these tools will only serve to replicate, and arguably exacerbate, unlawful or otherwise harmful outcomes at a faster rate and on a larger scale.’ Another piece in the symposium, also on AI decision support systems, argues that ‘it is important to carefully consider how these systems affect the conduct of hostilities, including through legal reviews, and to recognise the humanitarian consequences they pose for civilians through their use and experimentation.’

Other items:

• Event: CSET – Security and Emerging Technology Seminar Series– 10 April 2024: The Center for Security and Emerging technologies is organising an event titled ‘AI, Elections and Disinformation’ with CSET Research Fellow Josh Goldstein and a panel of outside experts. The event requires prior registration.