Microsoft has introduced a new service, called Correction, aimed at addressing a significant flaw in AI models, hallucinations, or factually incorrect responses. The tool identifies and revises erroneous AI-generated content by cross-referencing with accurate data sources, such as transcripts. Correction, available through Microsoft’s Azure AI Content Safety API, works with various models, including OpenAI’s GPT-4.
While Microsoft promotes Correction as a way to boost AI reliability, experts remain skeptical. Researchers warn that hallucinations are deeply ingrained in how AI models operate. Since these systems rely on statistical patterns rather than actual knowledge, completely eliminating false outputs might be impossible. They also caution that this solution could create new issues, like giving users a false sense of trust in AI outputs.
Despite these concerns, Microsoft is pushing to demonstrate the value of its AI tools, having invested billions in the technology. However, concerns about performance and cost are mounting, with some clients already pausing AI deployments due to inaccuracies and high expenses. Experts argue that AI, still in its developmental stages, is being rushed into industries without fully addressing its flaws.
Oman has launched the National Programme for Artificial Intelligence (AI) and Advanced Digital Technologies, aiming to secure global digital leadership and enhance the role of information and communications technology (ICT) in the economy. That initiative, running from 2024 to 2026, is part of the broader Oman Vision 2040 objectives, which seek to align the country with global advancements in AI.
The National Programme for Artificial Intelligence (AI) in Oman aims to establish a national open data platform to support entrepreneurs and decision-makers, create a research centre for AI, and develop a collaborative AI studio for technological solutions. It will also produce an Omani language model using generative AI and establish a centre for Fourth Industrial Revolution technologies in partnership with the World Economic Forum, alongside quantum computing initiatives.
Minister H E Saeed bin Hamood al Mawali emphasised the government’s commitment to integrating AI across sectors to enhance the digital economy’s contribution to GDP from 2% in 2021 to 10% by 2040. Additionally, the programme includes the AI Makers Initiative, opportunities for startups, and AI and data analysis training for Omani youth to prepare them as future innovators and creators.
Israel is proactively shaping its AI landscape by establishing a national expert forum on AI policy and regulation. Led by the Ministry of Innovation, Science, and Technology, this initiative demonstrates the government’s commitment to responsibly harnessing AI and unites experts to address its challenges and opportunities.
Its primary goal is to develop a comprehensive strategy for AI’s safe and ethical use, fostering innovation and maximising social and economic benefits. Specifically, by addressing data privacy, algorithmic bias, and ethical considerations, the forum aims to create a regulatory framework that supports technological advancements while protecting public interests.
Moreover, regular meetings will foster collaboration and dialogue among members, providing valuable insights to help the ministry navigate the complexities and risks of AI deployment. This collaborative environment will encourage sharing best practices and lessons learned from international AI governance experiences. Additionally, members will play a crucial role in shaping Israel’s policies in response to global developments in AI.
By actively monitoring international trends, the forum will ensure that Israel remains a leader in AI research and application while adhering to responsible governance practices. Ultimately, carefully selecting diverse and expert members will enrich discussions, thereby creating a comprehensive approach to AI governance that benefits all sectors of society.
OECD and the UN have established a significant partnership to enhance global governance for AI. That collaboration recognises the rapid advancement of AI technology and highlights the necessity for cohesive policy ecosystems to tackle its associated opportunities and challenges effectively.
By joining forces, the two organisations aim to provide comprehensive support to member states, ensuring that AI governance evolves with technological progress. The partnership underscores the importance of engaging diverse stakeholders, including scientists and academic institutions, to create a robust framework that addresses the complexities of AI governance. This inclusive approach guarantees that policy responses are well-informed and reflect the multifaceted impact of AI on society.
Furthermore, the OECD and UN will conduct regular assessments of AI-related risks and opportunities to ensure their initiatives remain aligned with the ever-evolving AI landscape. The proactive strategy will empower member states to navigate the intricacies of AI governance while fostering an inclusive approach that addresses both the benefits and challenges.
They share a vision of human-centred AI governance that prioritises safety, security, and trustworthiness, placing human interests at the forefront. This commitment to responsible governance is crucial for effectively managing the risks and disruptions posed by advancing AI technologies.
Why does this matter?
Ultimately, this collaboration emphasises the importance of rigorous scientific assessments and the role of the OECD’s AI Policy Observatory in understanding AI’s implications and fostering responsible innovation. The OECD and UN aim to equip countries with the necessary tools to harness AI’s potential while mitigating its risks, thereby promoting a globally inclusive approach to AI policy.
Generate: Biomedicines has entered a new partnership with Novartis to develop advanced protein therapies. Novartis is committing $65 million upfront, including $50 million in cash and $15 million in equity, with future milestone payments potentially exceeding $1 billion.
The partnership mirrors similar deals, including Generate’s ongoing work with Amgen, focusing on protein-based therapeutics such as antibodies, enzymes, and cytokines. AI-driven biotech firms are increasingly sought after by large pharmaceutical companies, with Novartis also striking deals with other AI-focused players like Alphabet’s Isomorphic Labs.
Led by CEO Mike Nally since 2021, Generate has rapidly expanded, raising over $600 million and pushing nearly 20 drug programs forward. The Novartis collaboration will enable Generate to further leverage its AI-based molecular design technologies, potentially filing up to six new drug applications within 18 months.
Generate’s financial position is strong, with over $350 million in cash reserves. CEO Nally stated that while an IPO could be an option, the company is well-positioned to continue without needing to go public, backed by its solid group of investors.
Tech giants Google and Nvidia are increasing their focus on AI in India. The companies made the announcement following a meeting with Prime Minister Narendra Modi during his recent US visit, signalling plans for deeper involvement in AI development across the country.
Nvidia CEO Jensen Huang commended Prime Minister Modi’s keen interest in AI and its potential to transform industries. Huang also recognised growing tech talent and startup ecosystem, which he views as vital to AI innovation and development. Nvidia has formed various partnerships in India to help scale AI capabilities.
Google CEO Sundar Pichai emphasised the Prime Minister’s push for India to adopt AI across sectors, including healthcare, education, and agriculture. Pichai confirmed Google’s intention to continue robust investments in AI to help further India’s digital transformation and ensure AI benefits the wider population.
Both Google and Nvidia are committed to supporting India’s technological growth, recognising its pivotal role in shaping the future of AI globally. Their collaboration with the Indian government aims to boost innovation and provide advanced AI solutions for various industries.
In mid-August, TSMC sealed a significant deal to acquire the AP8 facility from panel manufacturer Innolux in southern Taiwan, aiming to bolster its advanced packaging capacity. This new fab is set to produce advanced 3D Chip on Wafer on Substrate (CoWoS) IC packaging services, which are crucial for meeting the rising demand from AI server manufacturers. Expected to begin production in the latter half of 2025, the AP8 facility will have nine times the capacity of TSMC’s existing AP6 fab.
The acquisition, finalised for NTD 17.14 billion, came in under market expectations and was motivated by the need to bypass lengthy environmental assessments associated with new builds. TSMC intends to make internal modifications to the facility, enabling faster equipment installation, with deliveries set to begin as early as April next year.
TSMC Chairman C.C. Wei is optimistic about doubling CoWoS capacity in 2024 and 2025, with a goal of achieving a supply-demand balance by 2026. Analysts forecast that TSMC’s monthly CoWoS output could increase from 32,000 wafers this year to approximately 70,000 by the end of 2025, driven by this strategic expansion.
With an anticipated compound annual growth rate of over 50% for CoWoS capacity from 2022 to 2026, TSMC is committed to accelerating its fab construction timelines, aiming to reduce the typical 3-to-5 years down to just 2 years, ensuring it meets the surging market demand efficiently.
Bank of Canada Governor Tiff Macklem has said that businesses adopting AI could create short-term inflationary pressures by increasing demand. Speaking at an AI conference in Toronto, he noted that while AI-driven productivity growth may benefit the economy in the long run, its immediate effects could add to inflation.
Macklem pointed to rising electricity demand from new data centres as an example of AI’s economic impact. He emphasised that central bankers are working to understand how AI will affect the economy, inflation, and employment. Despite concerns, there is no current evidence that AI displaces labour at a level that would significantly impact overall employment rates.
The Bank of Canada has begun using AI to improve economic forecasting, data analysis, and efficiency, though its application is still in the early stages. Macklem compared the central bank’s cautious approach to AI to cautiously entering a dark room, feeling the way forward before making decisions.
Canada has also introduced a Voluntary Code of Conduct for the responsible development of generative AI systems. Macklem underlined that while AI offers potential benefits, it brings challenges and uncertainties that require careful management by policymakers.
At Amsterdam’s Vrije Universiteit, assistant professor Filip Ilievski is exploring the intriguing intersection of AI and human reasoning. His research focuses on enhancing AI’s ability to solve logic puzzles and riddles, revealing a significant gap in AI’s common sense reasoning. While AI excels at recognising patterns, it struggles with abstract thinking and temporal reasoning, as demonstrated in a recent study where OpenAI’s GPT-4 failed to determine if a character named Mable was alive at noon based on provided details.
Ilievski’s work suggests that researchers can unlock secrets about our cognitive processes by examining how both AI and humans tackle complex tasks. For instance, humans often trust their intuition, which can lead to errors, while AI can sometimes excel at problems that trip us up, such as the classic bat and ball riddle. However, when faced with novel puzzles, AI models still lag behind human performance, indicating a unique advantage in human reasoning skills.
Recent advancements show promise, as newer AI models like GPT-o1 have improved significantly, successfully answering challenging questions and demonstrating a better grasp of logic. Despite these strides, researchers caution that understanding AI may not directly illuminate how human cognition works, highlighting that the two systems operate differently. Ultimately, the collaboration between AI and human intellect could improve cognitive understanding in both realms.
Speaking at the UN Summit of the Future 2024, Google CEO Sundar Pichai described AI as the most transformative technology yet and announced a $120 million Global AI Opportunity Fund. The fund would provide AI education and training worldwide through partnerships with local NGOs and nonprofits.
Pichai highlighted four key areas where AI can contribute to sustainable development: language accessibility, scientific discovery, climate disaster alerts, and economic progress. He stressed the importance of harnessing AI for global advancement while addressing its risks.
He also warned of the potential for an ‘AI divide,’ where some regions may need to catch up in access to the technology. To combat this, Pichai called for smart global regulations that mitigate harm without promoting national protectionism, which could limit the benefits of AI.
Although Pichai did not mention the environmental impacts of AI, he emphasised the need for balanced regulation to ensure equal access and opportunities for AI development worldwide.
