ai Archives | Page 276 of 374 | Digital Watch Observatory

Experts at IGF 2024 address the dual role of AI in elections, emphasising empowerment and challenges

At IGF 2024, panellists explored AI’s role in elections, its potential for both empowerment and disruption, and the challenges it poses to democratic processes. Moderator Tapani Tarvainen led the discussion with contributions from Ayobangira Safari Nshuti, Roxana Radu, Babu Ram Aryal, and other experts.

Speakers noted that AI had been primarily used for self-promotion in campaigns, helping smaller candidates compete with limited resources. Roxana Radu highlighted AI’s positive role in voter outreach in India but warned of risks such as disinformation and public opinion manipulation. Ayobangira Safari Nshuti pointed to algorithmic biases and transparency issues in platforms as critical concerns, emphasising a recent case in Romania where AI-enabled manipulation caused election disruption.

Accountability of social media platforms became a focal point. Platforms increasingly rely on AI for content moderation, but their effectiveness in languages with limited online presence remains inadequate. Babu Ram Aryal stressed the need for stronger oversight, particularly in multilingual nations, while Dennis Redeker underscored the challenges of balancing regulation with free speech.

Panellists called for holistic solutions to safeguard democracy. Suggestions included enhancing platform transparency, implementing robust digital literacy programmes, and addressing social factors like poverty that exacerbate misinformation. Nana, an AI ethics specialist, advocated for proactive governance to adapt electoral institutions to technological realities.

The session concluded with a recognition that AI’s role in elections will continue to evolve. Panellists urged collaborative efforts between governments, civil society, and technology companies to ensure election integrity and maintain public trust in democratic systems.

Basis lands $34 million to revolutionise accounting

Basis, an AI startup, has secured $34 million in a Series A funding round to develop its AI-powered accounting automation product. The round, led by Khosla Ventures, attracted a diverse group of investors, including NFDG (the AI-focused fund managed by former GitHub CEO Nat Friedman and ex-Apple executive Daniel Gross), OpenAI board members Larry Summers and Adam D’Angelo, and Google’s chief scientist Jeff Dean.

The New York-based company is part of a growing group of AI startups creating autonomous agents—systems capable of performing tasks independently. Basis’ product, designed specifically for accounting firms, can handle various workflows such as entering transactions, verifying data accuracy, and integrating with popular ledger systems like QuickBooks and Xero. The product has already shown promising results, with large firms like Wiss reporting a 30% reduction in time spent on manual accounting tasks. Basis functions similarly to a junior accountant, allowing staff to focus on reviewing the AI’s work rather than completing tasks themselves.

Basis also aims to address the critical shortage of accountants in the US, exacerbated by retiring baby boomers and a decline in younger generations entering the profession. According to the Bureau of Labor Statistics, the accounting sector employs over 3 million people, but the number of candidates sitting for the CPA exam has fallen by 33% between 2016 and 2021. The shortage has led many firms to outsource work to countries like India. Moreover, with AI’s potential to automate tasks traditionally performed by accountants, the sector is expected to experience significant disruption. A 2023 OpenAI paper suggested that automation powered by large language models could eventually impact all accountant and auditor roles.

Partnership aims to advance AI in electric vehicles

Synopsys and SiMa.ai, two Silicon Valley-based companies, have announced a partnership to accelerate the development of energy-efficient AI chips designed for automotive applications. Synopsys, a leader in chip-design software, will collaborate with SiMa.ai, a startup known for its low-power hardware and software tailored for diverse AI functions.

The collaboration aims to meet the increasing demand for advanced AI technologies in electric vehicles, where efficient energy use is critical. SiMa.ai’s technology supports a range of applications, from driver-assistance systems that improve safety to voice assistants enabling hands-free commands. These tools often require different types of hardware, and the partnership allows automakers to simulate and select the best combinations for their needs.

The companies see this as a step towards integrating features like voice assistants into cars within the next three years. SiMa.ai’s CEO, Krishna Rangasayee, highlighted the importance of adapting data centre-level AI performance into power-efficient solutions for vehicles, ensuring both high performance and minimal energy consumption.

Inclusive AI governance: Perspectives from the Global South

At the 2024 Internet Governance Forum (IGF) in Riyadh, the Data and AI Governance coalition convened a panel to explore the challenges and opportunities of AI governance from the perspective of the Global South. The discussion delved into AI’s impacts on human rights, democracy, and economic development, emphasising the need for inclusive and region-specific frameworks.

Towards inclusive frameworks

Ahmad Bhinder, representing the Digital Cooperation Organization, stressed the importance of regional AI strategies. He highlighted the development of a self-assessment tool for AI readiness, designed to guide member states in governance and capacity development.

Similarly, Melody Musoni, Policy Officer at ECDPM, pointed to the African Union’s continental strategy as a promising example of unified AI governance. Elise Racine’s (Doctoral candidate at the University of Oxford) proposal for reparative algorithmic impact assessments underscored the need to address historical inequities, providing a blueprint for more equitable AI systems.

Ethics, rights, and regional challenges

The ethical dimensions of AI took centre stage, with Bianca Kremer, a member of the board of CGI.br and a professor at FGV Law School Rio, highlighting algorithmic bias in Brazil, where ‘90.5% of those arrested through facial recognition technologies are black and brown.’ This stark statistic underscored the urgent need to mitigate AI-driven discrimination.

Guangyu Qiao Franco from Radboud University emphasised the underrepresentation of Global South nations in AI arms control discussions, advocating for an inclusive approach to global AI governance.

Labour, economy, and sustainability

The panel explored AI’s economic and environmental ramifications. Avantika Tewari, PhD candidate at the Center for Comparative Politics and Political Theory at Jawaharlal Nehru University in New Delhi, discussed the exploitation of digital labour in AI development, urging fair compensation for workers in the Global South.

Rachel Leach raised concerns about the environmental costs of AI technologies, including embodied carbon, and criticised the lack of sustainability measures in current AI development paradigms.

Regional and global collaboration

Speakers highlighted the necessity of cross-border cooperation. Sizwe Snail ka Mtuze and Rocco Saverino, PhD candidate at the Free University of Brussels, examined region-specific approaches in Africa and Latin America, stressing the importance of tailored frameworks.

Luca Belli’s (Professor at Vilo School, Director of the Center for Technology and Society) observations on Brazil revealed gaps between AI regulation and implementation, emphasising the need for pragmatic, context-sensitive policies.

Actionable pathways forward

The discussion concluded with several actionable recommendations: fostering inclusive AI governance frameworks, implementing reparative assessments, addressing environmental and labour impacts, and prioritising digital literacy and regional collaboration.

‘Inclusive governance is not just a moral imperative but a practical necessity,’ Bhinder remarked, encapsulating the panel’s call to action. The session underscored the critical need for global cooperation to ensure AI serves humanity equitably.

All transcripts from the Internet Governance Forum sessions can be found on dig.watch.

Parliamentary panel at IGF discusses ICTs and AI in counterterrorism efforts

At the 2024 Internet Governance Forum (IGF) in Riyadh, a panel of experts explored how parliaments can harness information and communication technologies (ICTs) and AI to combat terrorism while safeguarding human rights. The session, titled ‘Parliamentary Approaches to ICT and UN SC Resolution 1373,’ emphasised the dual nature of these technologies—as tools for both law enforcement and malicious actors—and highlighted the pivotal role of international collaboration.

Legislation and oversight in a digital era

David Alamos, Chief of the UNOCT programme on Parliamentary Engagement, set the stage by underscoring the responsibility of parliaments to translate international frameworks like UN Security Council Resolution 1373 into national laws. ‘Parliamentarians must allocate budgets and exercise oversight to ensure counterterrorism efforts are both effective and ethical,’ Alamos stated.

Akvile Giniotiene of the UN Office of Counterterrorism echoed this sentiment, emphasising the need for robust legal frameworks to empower law enforcement in leveraging new technologies responsibly.

Opportunities and risks in emerging technologies

Panelists examined the dual role of ICTs and AI in counterterrorism. Abdelouahab Yagoubi, a member of Algeria’s National Assembly, highlighted AI’s potential to enhance threat detection and predictive analysis.

At the same time, Jennifer Bramlette from the UN Counterterrorism Committee stressed the importance of digital literacy in fortifying societal resilience. On the other hand, Kamil Aydin and Emanuele Loperfido of the OSCE Parliamentary Assembly cautioned against the misuse of these technologies, pointing to risks like deepfakes and cybercrime-as-a-service, enabling terrorist propaganda and disinformation campaigns.

The case for collaboration

The session spotlighted the critical need for international cooperation and public-private partnerships to address the cross-border nature of terrorist threats. Giniotiene called for enhanced coordination mechanisms among nations, while Yagoubi praised the Parliamentary Assembly of the Mediterranean for fostering knowledge-sharing on AI’s implications.

‘No single entity can tackle this alone,’ Alamos remarked, advocating for UN-led capacity-building initiatives to support member states.

Balancing security with civil liberties

A recurring theme was the necessity of balancing counterterrorism measures with the protection of human rights. Loperfido warned against the overreach of security measures, noting that ethical considerations must guide the development and deployment of AI in law enforcement.

An audience query on the potential misuse of the term ‘terrorism’ further underscored the importance of safeguarding civil liberties within legislative frameworks.

Looking ahead

The panel concluded with actionable recommendations, including updating the UN Parliamentary Handbook on Resolution 1373, investing in digital literacy, and ensuring parliamentarians are well-versed in emerging technologies.

‘Adapting to the rapid pace of technological advancement while maintaining a steadfast commitment to the rule of law is paramount,’ Alamos said, encapsulating the session’s ethos. The discussion underscored the indispensable role of parliaments in shaping a global counterterrorism strategy that is both effective and equitable.

NeurIPS conference showcases AI’s rapid growth

The NeurIPS conference, AI’s premier annual gathering, drew over 16,000 computer scientists to British Columbia last week, highlighting the field’s rapid growth and transformation. Once an intimate meeting of academic outliers, the event has evolved into a showcase for technological breakthroughs and corporate ambitions, featuring major players like Alphabet, Meta, and Microsoft.

Industry luminaries like Ilya Sutskever and Fei-Fei Li discussed AI’s evolving challenges. Sutskever emphasised AI’s unpredictability as it learns to reason, while Li called for expanding beyond 2D internet data to develop “spatial intelligence.” The conference, delayed a day to avoid clashing with a Taylor Swift concert, underscored AI’s growing mainstream prominence.

Venture capitalists, sponsors, and tech giants flooded the event, reflecting AI’s lucrative appeal. The number of research papers accepted has surged tenfold in a decade, and discussions focused on tackling the costs and limitations of scaling AI models. Notable attendees included Meta’s Yann LeCun and Google DeepMind’s Jeff Dean, who advocated for ‘modular’ and ‘tangly’ AI architectures.

In a symbolic moment of AI’s widening reach, 10-year-old Harini Shravan became the youngest ever to have a paper accepted, illustrating how the field now embraces new generations and diverse ideas.

Meta enhances Ray-Ban smart glasses with AI video and translation

Meta Platforms has introduced significant upgrades to its Ray-Ban Meta smart glasses, adding AI video capabilities and real-time language translation. The updates, announced during Meta’s Connect conference in September, are now available through the v11 software rollout for Early Access Program members.

The new AI video feature lets the smart glasses process visuals and answer user queries in real-time. Additionally, the glasses can now translate speech between English and Spanish, French, or Italian, providing translations via open-ear speakers or as text on a connected phone.

Meta also integrated the Shazam music identification app into the glasses, enhancing their functionality for users in the US and Canada. Earlier AI upgrades, such as setting reminders and scanning QR codes via voice commands, continue to expand the glasses’ utility.

Revitalising trust with AI: Boosting governance and public services

AI is reshaping public governance, offering innovative ways to enhance services and restore trust in institutions. The discussion at the Internet Governance Forum (IGF) in Riyadh, moderated by Brandon Soloski of Meridian International, focused on using AI to streamline services like passport processing and tax systems, while also addressing privacy and data sovereignty concerns. Open-source AI was highlighted as a critical tool for democratising access and fostering innovation, particularly in developing nations.

Global regulatory frameworks were a central theme, with panellists underscoring the need for harmonisation to avoid fragmentation and ensure seamless interoperability across borders. Economist and policy analyst at the OECD, Lucia Russo, discussed regulatory approaches such as the EU AI Act, which aims to create a comprehensive legal framework. Brandon Soloski and Sarim Aziz from Meta pointed to the benefits of principle-based frameworks in other regions, which provide flexibility while maintaining oversight. Pellerin Matis, Vice President of Global Government Affairs at Oracle, emphasised the importance of public-private partnerships, which allow governments to leverage private sector expertise and startup innovation for effective AI implementation.

The panellists explored how AI can enhance public services, highlighting its role in healthcare, agriculture, and public safety. Examples included AI-driven tools that improve patient care and streamline food production. However, challenges like data protection, trust in AI systems, and the balance between innovation and regulation were also discussed. Anil Pura, an audience member from Nepal, contributed valuable perspectives on the need for education and transparency to foster public trust.

Transparency and education were recognised as fundamental for building trust in AI adoption. Panellists agreed that ensuring citizens understand how AI technologies work and how their data is protected is essential for encouraging adoption. They called for governments to work closely with civil society and academia to create awareness and promote responsible AI use.

The discussion concluded with a call to strengthen collaborations between governments, private companies, and startups. Brandon Soloski highlighted how partnerships could drive responsible AI innovation, while Pellerin Matis stressed the importance of ethical and regulatory considerations to guide development. The session ended on an optimistic note, with panellists agreeing on AI’s immense potential to improve government efficiency and enhance public trust.

All transcripts from the Internet Governance Forum sessions can be found on dig.watch.

IGF 2024 panellists highlight infrastructure, literacy, and fair digital access

The Internet Governance Forum 2024 (IGF) brought together global stakeholders to discuss the implementation of the Global Digital Compact (GDC), aiming to address digital inequalities and foster cross-sector partnerships. The session spotlighted key challenges such as funding gaps, cultural adaptation of digital initiatives, and sustainability concerns in infrastructure development.

Isabel De Sola from the Office of the Tech Envoy emphasised stakeholder collaboration and revealed plans for an upcoming GDC implementation roadmap. Roy Eriksson, Finland‘s Ambassador for Global Gateway, shared successes from AI strategy projects in African nations, illustrating how capacity-building partnerships can close technology gaps. Kevin Hernandez of the Universal Postal Union presented the Connect.Post programme, which aims to connect global post offices to digital networks by 2030.

Discussions also underscored energy efficiency and sustainability in digital infrastructure. Nandipha Ntshalbu highlighted the need to balance technological growth with environmental considerations. Data governance and cybersecurity frameworks were identified as critical, with Shamsher Mavin Chowdhury stressing the importance of inclusive frameworks to protect the interests of developing countries.

Innovative projects demonstrated local impact, such as Damilare Oydele’s Library Tracker for African libraries and Patricia Ainembabazi’s efforts promoting regional knowledge-sharing platforms. However, Alisa Heaver of the Dutch Ministry of Economic Affairs raised concerns about aligning GDC objectives with existing frameworks to avoid redundancy.

The IGF session concluded with a unified call for continued collaboration. Despite challenges, there was optimism that effective partnerships and targeted initiatives can ensure secure, inclusive, and sustainable digital progress worldwide.

Balancing innovation and oversight: AI’s future requires shared governance

At IGF 2024, day two in Riyadh, policymakers, tech experts, and corporate leaders discussed one of the most pressing dilemmas in the AI age: how to foster innovation in large-scale AI systems while ensuring ethical governance and regulation. The session ‘Researching at the frontier: Insights from the private sector in developing large-scale AI systems’ reflected the urgency of navigating AI’s transformative power without losing sight of privacy, fairness, and societal impact.

Ivana Bartoletti, Chief Privacy and AI Governance Officer at Wipro called on governments to better use existing privacy and data protection laws rather than rush into new AI-specific legislation. ‘AI doesn’t exist in isolation. Privacy laws, consumer rights, and anti-discrimination frameworks already apply,’ she said, stressing the need for ‘privacy by design’ to protect individual freedoms at every stage of AI development.

Basma Ammari from Meta added a private-sector perspective, advocating for a risk-based and principles-driven regulatory approach. Highlighting Meta’s open-source strategy for its large language models, Ammari explained, ‘More diverse global input strips biases and makes AI systems fairer and more representative.’ She added that collaboration, rather than heavy-handed regulation, is key to safeguarding innovation.

Another expert, Fuad Siddiqui, EY’s Emerging Tech Leader, introduced the concept of an ‘intelligence grid,’ likening AI infrastructure to electricity networks. He detailed AI’s potential to address real-world challenges, citing applications in agriculture and energy sectors that improve productivity and reduce environmental impacts. ‘AI must be embedded into resilient national strategies that balance innovation and sovereignty,’ Siddiqui noted.

Parliamentarians played a central role in the discussion, raising concerns about AI’s societal impacts, particularly on jobs and education. ‘Legislators face a steep learning curve in AI governance,’ remarked Silvia Dinica, a Romanian senator with a background in mathematics. Calls emerged for upskilling initiatives and AI-driven tools to support legislative processes, with private-sector partnerships seen as crucial to addressing workforce disruption.

The debate over AI regulation remains unsettled, but a consensus emerged on transparency, fairness, and accountability. Panelists urged parliamentarians to define national priorities, invest in research on algorithm validation, and work with private stakeholders to create adaptable governance frameworks. As Bartoletti aptly summarised, ‘The future of AI is not just technological—it’s about the values we choose to protect.’

All transcripts from the Internet Governance Forum sessions can be found on dig.watch.