In the first part of her address, the Secretary noted the opportunities and risks of AI generally, highlighting the use of automation in the financial sector for many years and the department’s regulation of it, citing model risk management guidance and third-party risk management. However, she warned of other risks to do with the ‘complexity and opacity’ of new AI systems. She noted the difficulty of regulating an increasing amount of AI systems, but also avoiding dependence on a few.
Why is this important?
To address this “rapidly evolving field”, the Treasury released a report providing current use cases and practices of AI for cybersecurity and fraud prevention in the financial sector. Work is also being done with international partners in this field, Yellen said. Towards the end of her address, the Treasury Secretary mentioned a request for comments from private and public actors, as well as a round table discussion after this. She also mentioned that the council’s monitoring will be continued and adapted in line with the rapid implementation of AI in the financial sector.
Elon Musk, the billionaire CEO of Tesla, SpaceX, and the social media company X announced on Monday that he would ban Apple devices from his companies if Apple integrates OpenAI technology at the operating system level. Musk called this move an ‘unacceptable security violation’ and declared that visitors would have to leave their Apple devices in a Faraday cage at the entrance to his facilities.
The statement followed Apple’s announcement of new AI features across its apps and operating platforms, including a partnership with OpenAI to incorporate ChatGPT technology into its devices. Apple emphasised that these AI features are designed with privacy at their core, using both on-device processing and cloud computing to ensure data security. Musk, however, expressed scepticism, arguing that Apple’s reliance on OpenAI undermines its ability to protect user privacy and security effectively.
If Apple integrates OpenAI at the OS level, then Apple devices will be banned at my companies. That is an unacceptable security violation.
Industry experts, such as Ben Bajarin, CEO of Creative Strategies, believe that Musk’s stance is unlikely to gain widespread support. Bajarin noted that Apple aims to reassure users that its private cloud services are as secure as on-device data storage. He explained that Apple anonymises and firewalls user data, ensuring that Apple itself does not access it.
Musk’s criticism of OpenAI is not new; he co-founded the organisation in 2015 but sued it earlier this year, alleging it strayed from its mission to develop AI for the benefit of humanity. Musk has since launched his own AI startup, xAI, valued at $24 billion after a recent funding round, to compete directly with OpenAI and develop alternatives to its popular ChatGPT.
Meta Platforms, the owner of Facebook, announced it is developing AI technology tailored specifically for Europe, taking into account the region’s linguistic, geographic, and cultural nuances. The company will train its large language models using publicly shared content from its platforms, including Instagram and Facebook, ensuring that private posts are excluded to maintain user privacy.
Last month, Meta revealed plans to inform Facebook and Instagram users in Europe and the UK about how their public information is utilised to enhance and develop AI technologies. The move aims to increase transparency and reassure users about data privacy.
By focusing on localised AI development, Meta hopes to serve the European market better, reflecting the region’s diverse characteristics in its technology offerings. That effort underscores Meta’s commitment to respecting user privacy while advancing its AI capabilities.
At Apple’s annual developer conference on Monday, the tech giant is anticipated to unveil how it’s integrating AI across its software suite. The integration includes updates to its Siri voice assistant and a potential collaboration with OpenAI, the owner of ChatGPT. With its reputation on the line, Apple aims to reassure investors that it remains competitive in the AI landscape, especially against rivals like Microsoft.
Apple faces the challenge of demonstrating the value of AI to its vast user base, many of whom are not tech enthusiasts. Analysts suggest that Apple needs to showcase how AI can enhance user experiences, a shift from its previous emphasis on enterprise applications. Despite using AI behind the scenes for years, Apple has been reserved in highlighting its role in device functionality, unlike Microsoft’s more vocal approach with OpenAI.
The spotlight is on Siri’s makeover, which is expected to enable more seamless control over various apps. Apple aims to make Siri smarter by integrating generative AI, potentially through a partnership with OpenAI. The move is anticipated to improve user interactions with Siri across different apps, enhancing its usability and effectiveness. Also, Apple recently introduced an AI-focused chip in its latest iPad Pro models, signalling its commitment to AI development. Analysts predict that Apple will provide developers with insights into leveraging these capabilities to support AI computing. Additionally, reports suggest Apple may discuss its plans for using its chips in data centres, which could enhance cloud computing capabilities while maintaining privacy and security features.
The Apple Worldwide Developers Conference (WWDC 2024) will run until Friday, offering developers insights into app updates and new tools. Investors are hopeful that Apple’s AI advancements will drive sales of new iPhones and boost the company’s competitive edge amid fierce global competition.
Three US lawmakers have raised concerns about NewsBreak, a popular news aggregation app, due to its Chinese origins and use of AI tools that have produced erroneous stories. Senator Mark Warner, chair of the Intelligence Committee, emphasised the threat posed by technologies from adversarial countries. At the same time, Representative Raja Krishnamoorthi highlighted the need for transparency regarding any ties to the Chinese Communist Party (CCP). Representative Elise Stefanik pointed to the backing by IDG Capital, a Beijing-based private equity firm, as a reason for increased scrutiny.
NewsBreak, launched in the US in 2015, was originally a subsidiary of the Chinese news app Yidian, founded by Jeff Zheng. Despite being labelled an American company by its spokesperson, court documents and other evidence reveal historical links to Chinese investors and engineers based in China. Notably, Yidian has received praise from Chinese Communist Party officials for disseminating government propaganda, although there is no evidence that NewsBreak has censored or produced pro-China news.
The primary investors in NewsBreak include San Francisco-based Francisco Partners and Beijing-based IDG Capital. IDG Capital, which the Pentagon has listed as allegedly working with Beijing’s military, denies any such association. Francisco Partners has described the scrutiny as ‘false and misleading,’ but the lawmakers maintain their stance on carefully examining the app’s potential risks to US interests.
Google has issued new guidance for developers building AI apps distributed through Google Play in response to growing concerns over the proliferation of AI-powered apps designed to create deepfake nude images. The platform recently announced a crackdown on such applications, signalling a firm stance against the misuse of AI for generating non-consensual and potentially harmful content.
The move comes in the wake of alarming reports highlighting the ease with which these apps can manipulate photos to create realistic yet fabricated nude images of individuals. Reports have surfaced about apps like ‘DeepNude’ and its clones, which can strip clothes from images of women to produce highly realistic nude photos. Another report detailed the widespread availability of apps that could generate deepfake videos, leading to significant privacy invasions and the potential for harassment and blackmail.
Apps offering AI features have to be ‘rigorously tested’ to safeguard against prompts that generate restricted content and have to provide a way for users to signal it. Google strongly suggests that developers document the recommended tests before launching them, as Google could ask them to be reviewed in the future. Additionally, developers can’t advertise that their app breaks any of Google Play’s rules at the risk of getting banned from the app store. The company is also publishing other resources and best practices, like its People + AI Guidebook, which aims to support developers building AI apps.
Why Does It Matter?
The proliferation of AI-driven deepfake apps on platforms like Google Play undermine personal privacy and consent by allowing anyone to generate highly realistic and often explicit content of individuals without their knowledge or consent. Such misuse can lead to severe reputational damage, harassment, and even extortion, affecting both individuals and public figures alike.
A new mobile camera unit will be deployed in UK’s East Yorkshire and North Lincolnshire to catch drivers using mobile phones and those not wearing seat belts. In partnership with National Highways, Safer Roads Humber will operate the AI-equipped camera for a week starting Monday, 10 June. The AI technology identifies potential lawbreakers, with images reviewed by an officer to confirm violations before prosecution.
Offenders face significant penalties: a £200 fine and six points on their licence for using a handheld phone while driving, and a £100 fine for not wearing a seat belt. Drivers are also responsible for ensuring passengers under 14 are belted in. Sometimes, offenders may be offered an educational course instead of prosecution.
Ian Robertson from Safer Roads Humber highlighted the enhanced enforcement capabilities of this new equipment. While their current safety camera vans already detect such offences, the advanced AI technology of the new unit provides added capacity to improve road safety.
Massachusetts Institute of Technology (MIT) researchers have developed an AI chatbot that engages users in conversations with their future selves, aiming to promote thoughtful decision-making on health, finances, and career paths. By simulating an individual’s older self and drawing from synthetic memories, ‘Future You’ encourages users to consider their long-term aspirations and make informed choices in the present.
The process towards creating your chatbot has two main components. First, users initiate conversations by providing personal information, then, they upload a portrait image, digitally aged to depict their future selves. The chatbot utilises this data to generate coherent responses, leveraging OpenAI’s GPT3.5 language model to facilitate meaningful exchanges. The user’s new generated persona can give advice and reflect on their life to provide insights into the user’s future.
Why is this chatbot useful?
Users are told the ‘future self’ is not a prediction but rather a potential future self based on the information they provided. They are encouraged to explore different futures by changing their answers to the questionnaire. After several conversations with his ‘future self’, MIT Media Lab’s Pat Pataranutaporn reported his most profound discussion was when the chatbot reminded him to spend time with his parents, as they would not be around forever.
In a preprint scientific paper, the MIT project has shown promise in alleviating anxiety and fostering a stronger connection to one’s future self, empowering individuals to pursue specific goals, adopt healthier habits, and plan for their financial future. Behavioural science experts commend the project’s innovative approach, emphasising the potential of such tools to influence beneficial decision-making by making the future self more salient and relevant to the present.
Later this month, Berlin will see the launch of the world’s first cyber brothel, offering customers the opportunity to book time with AI sex dolls. The new service, spearheaded by Cybrothel founder Philipp Fussenegger, allows users to interact verbally and physically with the AI dolls, catering to a growing demand for more interactive AI experiences in the adult entertainment industry.
Generative AI is increasingly being integrated into the adult entertainment sector. A report by SplitMetrics shows that AI companion apps have been downloaded 225 million times on the Google Play Store, indicating a lucrative market. These AI companions often charge fees and collect user data, which is frequently shared with third parties, raising privacy concerns.
Experts have voiced significant concerns about the potential harms of merging AI with adult entertainment. Issues include the reinforcement of gender stereotypes, addiction risks, and privacy violations. AI chatbots, according to Mozilla’s privacy researcher Misha Rykov, target lonely individuals and can exacerbate mental health challenges. Furthermore, content warnings for themes of abuse, violence, and underage relationships have been added to several AI chatbots by Mozilla.
Despite these concerns, some industry leaders argue that AI can enhance the sexual experience without replacing human interaction. Philipp Hamburger from Lovehoney emphasises AI’s role in ethically improving user experience. Additionally, Ruben Cruz from The Clueless Agency believes AI can help mitigate ethical issues by preventing the explicit sexualisation of real individuals in adult content. However, the broader impact on real-world relationships and the potential for harmful assumptions about consent remain critical issues that need addressing.
China’s ByteDance, the parent company of TikTok, plans to invest around $2.13 billion to establish an AI hub in Malaysia. The plan includes an additional $320 million to expand data centre facilities in Johor state, according to Malaysia’s Trade Minister Tengku Zafrul Aziz.
The development follows significant investments by other tech giants in Malaysia. Google recently announced a $2 billion investment to create its first data centre and Google Cloud region in the country, while Microsoft is set to invest $2.2 billion to enhance cloud and AI services.
The investment is expected to boost Malaysia’s digital economy, aiming to increase its contribution to 22.6% of its GDP by 2025, underscoring the county’s growing importance as a digital economy hub in Southeast Asia.
