OpenAI Archives | Page 9 of 49 | Digital Watch Observatory

OpenAI set to spend $10bn on Broadcom AI chips

OpenAI has reportedly placed a $10bn order with Broadcom to mass-produce custom AI chips, due for shipment in 2026. Sources told the Financial Times that the move would help reduce OpenAI’s dependence on Nvidia, its primary supplier.

Sam Altman recently said OpenAI will use ‘well over 1m GPUs’ by the end of 2025, highlighting the company’s accelerating demand for computing power. In contrast, Elon Musk’s xAI is expected to double its Nvidia Hopper GPUs to around 200,000.

Broadcom confirmed a large custom chip order during its latest earnings call, without naming the buyer. The company’s AI revenue rose 63 percent to $5.2bn, chip sales grew 57 percent to $9.1bn, and shares gained nearly 5 percent.

The new order is expected to be for internal use rather than external customers. Industry observers suggest that OpenAI’s decision signals a strategic shift, allowing the ChatGPT maker to secure supply for its AI expansion while diversifying beyond Nvidia.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

New OpenAI platform aims to connect employers and talent

OpenAI has announced plans to launch an AI-powered hiring platform to compete with LinkedIn directly. The service, OpenAI Jobs Platform, is expected to debut by mid-2026.

Applications CEO Fidji Simo said the platform will help businesses and employees find ideal matches using AI, with tailored options for small businesses and local governments. The Texas Association of Business plans to use the platform to connect employers with talent.

The move highlights OpenAI’s efforts to expand beyond ChatGPT into a broader range of applications, including a browser, a social media app, and recruitment. The company faces intense competition from Microsoft-owned LinkedIn, which has been adding AI features of its own.

Alongside the hiring initiative, OpenAI is preparing to pilot its Certifications programme through the OpenAI Academy. The scheme will provide certificates for AI proficiency, with Walmart among the first partners.

OpenAI aims to certify 10 million Americans by 2030 as part of its commitment to advancing AI literacy.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

GPT-5 flunks Kindergarten test despite PhD-level promise

Critics quickly derided OpenAI’s newly released GPT-5 for failing tasks that a five-year-old could ace, raising questions about the disparity between hype and performance.

Despite being promoted as ‘PhD-level’, the model produced a distorted, blob-like map of North America and invented mismatched portraits of US presidents with fictional names.

AI researcher Gary Marcus lowered the threshold by giving GPT-5 a kindergarten-level challenge. The result was a clear fail. He posted: ‘GPT-5 failed a kindergarten-level task. Speechless.’ He criticised the rushed rollout and the hype that may have obscured the model’s visual reasoning weaknesses.

Further tests exposed inconsistencies: when asked to map France and label its 12 most populous cities, GPT-5 returned inaccurate or incomplete results, omitting Paris entirely and naming Orléans despite its lower ranking.

Oddly, when the same queries were posed in text-only form, the model performed better, highlighting the weakness in its image generation and visual logic.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Weekly #228 Roadmap to the digital future: WSIS+20 zero draft paves the way

29 August – 5 September 2025

Person, Adult, Male, Man, Book, Comics, Publication, Face, Head

Dear readers,

In 2003-2005, a landmark UN summit – the World Summit on the Information Society (WSIS) – outlined a vision for an inclusive information society, set out recommendations for making this vision a reality, and laid the basis for much of what we call today the global digital governance architecture. Twenty years later, UN member states are looking at progress made in achieving the goals set back then and areas requiring further effort, as well as at whether the WSIS architecture needs updates. This unfolding WSIS+20 review process will end in December 2025 with a high-level meeting of the UN General Assembly. In the lead-up to the meeting, negotiations and consultations are now focused on concrete text for what will become a WSIS+20 outcome document. This concrete text – called the zero draft – was released last week.

Digital divides and inclusion take centre stage in the zero draft. While connectivity has expanded – 95% of the global population is now within reach of broadband, and internet use has grown from 15% in 2005 to 67% in 2025 – significant gaps remain. Disparities persist across countries, urban and rural areas, genders, persons with disabilities, older populations, and minority language speakers. The draft calls for affordable entry-level broadband, local multilingual content, digital literacy, and mechanisms to connect the unconnected, ensuring equitable access.

The digital economy continues to transform trade, finance, and industry, creating opportunities for small and women-led businesses but also risks deepening inequalities through concentrated technological power and automation. Against this backdrop, the draft outlines a commitment to supporting the development of digital financial services, and a call for stakeholders to foster ‘open, fair, inclusive and non-discriminatory digital environments.

Environmental sustainability is a key consideration, as ICTs facilitate monitoring of climate change and resource management, yet their growth contributes to energy demand, emissions, and electronic waste. Standing out in the draft is a call for the development of global reporting standards on environmental impacts, and of global standards for sustainable product design, and circular economy practices to align digital innovation with environmental goals.

The Zero Draft reaffirms human rights, confidence and security, and multistakeholder internet governance as central pillars of the digital ecosystem. Human rights are positioned as the foundation of digital cooperation, with commitments to protect freedom of expression, privacy, access to information, and the rights of women, children, and other vulnerable groups. Strengthening confidence and security in the use of technology is seen as essential for innovation and sustainable development, with emphasis on protecting users from threats such as online abuse and violence, hate speech, and misinformation, while ensuring safeguards for privacy and freedom of expression.

The draft outlines a series of key (desirable) attributes for the internet – open, free, global, interoperable, reliable, secure, stable – and highlights the need for more inclusive internet governance discussions, across stakeholder groups (governments, the private sector, civil society, academia, and technical communities) and across developed and developing countries alike.

To advance capacity building in relation to AI, the draft proposes a UN AI research programme and AI capacity building fellowship, both with a focus on developing countries. In parallel, the draft welcomes ongoing initiatives such as the Independent International Scientific Panel on AI and the Global Dialogue on AI Governance.

Recognising the critical importance of global cooperation in internet governance, the draft designates the Internet Governance Forum (IGF) as a permanent UN body and calls for enhanced secretariat support, enhanced working methods, and reporting on outcomes to UN entities and processes (which are then called to duly take these outputs into account in their work). The long-discussed issue of IGF financial sustainability is addressed in the form of a request for the UN Secretary-General to make proposals on future funding.

Finally, the draft looks at the interplay between WSIS, the Global Digital Compact and the 2030 Agenda for Sustainable Development, and outlines several mechanisms for better connecting them and avoiding duplication and overlaps. These include a joint WSIS-GDC implementation roadmap, the inclusion of GDC review and follow-up into existing annual WSIS mechanisms (at the level of the Commission on Science and Technology for Development and the Economic and Social Council), and reviews in GDC-WSIS alignments at the GA level Speaking of overall reviews, the draft also envisions a combined review of Agenda 2030 and of outcomes of the WSIS-GDC joint implementation roadmap in 2030, as well as a WSIS+30 review in 2035.

Looking ahead

The Zero Draft sets the stage for intense negotiations ahead of the December 2025 High-Level Meeting. Member states and other stakeholders are invited to submit comments until 26 September. It then remains to be seen what a second version of the outcome document will look like, and which elements are kept, revised, or removed.👉 Follow the process with us on our dedicated web page, where we will track key developments, highlight emerging debates, and provide expert analysis as the negotiations unfold.

DW Team

Highlights from the week of 29 August – 5 September 2025

Switzerland launches Apertus, an open multilingual AI model

The model is designed to boost innovation while remaining fully transparent and accessible to all.

TSMC Semiconductors AI A14 logic process

TSMC faces curbs on shipping US tech to China

The USA revoked TSMC’s licence to ship advanced technology to China, adding pressure to global semiconductor supply chains.

Google AI Mode Search Labs restaurant reservations US EU

Google avoids forced breakup in search monopoly trial

The ruling bars Google from exclusive distribution deals for products like Search, Chrome and Gemini.

CJEU dismisses bid to annul EU-US data privacy framework

The EU General Court upheld the EU–US Data Privacy Framework, rejecting claims it lacks adequate safeguards and independence in oversight of US data practices involving personal data from the EU.

France and Germany announce a joint Economic Agenda

The document flagship outlines projects in AI, quantum, cloud, and space, promotes a Franco-German digital ecosystem for public services, and sets the stage for the 2025 European Digital Sovereignty Summit.

Salt Typhoon espionage campaign revealed through global cybersecurity advisory

A joint cybersecurity advisory details how Salt Typhoon exploited unpatched network-edge devices to infiltrate telecommunications, military and government systems across 13 countries.

SCO Tianjin Summit puts tech, AI, and digital governance high on the agenda

The SCO Tianjin Declaration emphasised cyber sovereignty, inclusive AI development, global cybersecurity norms, and stronger cooperation in the digital economy.

Nigeria drafts framework for AI use in governance and services

A global report found 63% of employers say AI has significantly boosted productivity at work.

View all updates

READING CORNER

BLOG featured image 2025 From summer disillusionment to autumn clarity

From summer disillusionment to autumn clarity: Ten lessons for AI – Diplo

As classrooms and negotiation tables fill again, a pressing question lingers: where is AI headed? This summer marked a turning point, as the ‘bigger is better’ narrative faltered. This blog captures ten key lessons from a season of AI disillusionment.

Why apprenticeship and storytelling are the future of learning in the AI Era – Diplo

Why apprenticeship and storytelling are the future of learning in the AI Era AI is forcing us to ask a deeper question: what is the real purpose of learning?

UPCOMING EVENTS

Diplo EVENT 2025 African Priorities for the Global Digital

African priorities for the Global Digital Compact – Diplo Event

The webinar will bring together African experts from technology, development, diplomacy and policy domains to discuss which digital issues must be urgently prioritised to keep Africa on course in a rapidly changing world.

UN Cyber Dialogue 2025 web banners 678x248px 1

UN Cyber Dialogue: Shaping the future global mechanism for cyber-stability – Diplo Event

In this one-hour session, several experts – Asoke Mukerji, Isaac Morales Tenorio, and Fan Yang will debate the future of global cyber negotiations — tackling obstacles, testing new ideas, and asking whether the UN dialogue can move from compromise to real progress.

Human Rights Council (HRC) 60th session

This session provides a key platform for the international community to discuss, promote, and protect human rights worldwide.

Decoding the UN CSTD Working Group on Data Governance online series

The event will discuss the progress made by the Multi-Stakeholder Working Group on Data Governance and the expectations for the next meeting, which will take place on 15-16 September.

DeepSeek prepares new AI agent model to rival US competitors

According to people familiar with the plans, Chinese startup DeepSeek is developing an AI model with enhanced agent features to compete with US firms such as OpenAI.

The Hangzhou-based company intends for the system to perform multi-step tasks with limited input and adapt from its previous actions.

Founder Liang Wenfeng has urged his team to prepare the release before the end of 2025. The project follows DeepSeek’s earlier success with R1, a reasoning-focused model launched in January that attracted attention for its low development costs.

Since then, DeepSeek has delivered only incremental updates while rivals in China and the US have accelerated new product launches.

The shift towards AI agents reflects a broader industry move to develop tools capable of managing complex real-world tasks, from research to coding, with less reliance on users. OpenAI, Anthropic, Microsoft, and Manus AI have already introduced similar projects.

Most systems still require significant oversight, highlighting the challenges of building fully autonomous agents.

DeepSeek declined to comment on the development.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Key AI researchers depart Apple for rivals Meta and OpenAI

Apple is confronting a significant exodus of AI talent, with key researchers departing for rival firms instead of advancing projects in-house.

The company lost its lead robotics researcher, Jian Zhang, to Meta’s Robotics Studio, alongside several core Foundation Models team members responsible for the Apple Intelligence platform. The brain drain has triggered internal concerns about Apple’s strategic direction and declining staff morale.

Instead of relying entirely on its own systems, Apple is reportedly considering a shift towards using external AI models. The departures include experts like Ruoming Pang, who accepted a multi-year package from Meta reportedly worth $200 million.

Other AI researchers are set to join leading firms like OpenAI and Anthropic, highlighting a fierce industry-wide battle for specialised expertise.

At the centre of the talent war is Meta CEO Mark Zuckerberg, offering lucrative packages worth up to $100 million to secure leading researchers for Meta’s ambitious AI and robotics initiatives.

The aggressive recruitment strategy is strengthening Meta’s capabilities while simultaneously weakening the internal development efforts of competitors like Apple.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Statsig to be acquired by OpenAI in £1.1bn deal

OpenAI has agreed to acquire the product development startup Statsig in an all-stock deal valued at £1.1 billion, marking one of the most significant acquisitions in the company’s history.

Statsig, founded in 2021, provides tools for developers to test and manage new features. Upon completion of the deal, Statsig’s founder and CEO, Vijaye Raji, will join OpenAI as the new chief technology officer (CTO) for applications.

Raji will report to OpenAI Applications CEO Fidji Simo and lead product engineering for key products such as ChatGPT.

The acquisition is part of a broader trend of significant deals for the AI company this year, which recently concluded a £6.5 billion all-stock acquisition of an AI device startup. OpenAI’s expanding valuation, which reached £300 billion following a March funding round, has supported this growth.

The company is reportedly discussing a further share sale that could increase its valuation to £500 billion. The completion of the Statsig deal is subject to regulatory approval, after which the company will continue to operate independently from its Seattle office, with its employees joining the OpenAI team.

Other leadership changes at OpenAI include the appointment of Srinivas Narayanan as CTO for B2B applications and Kevin Weil’s move to a new team focused on AI for Science.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

ChatGPT hit by widespread outage: ‘Our work partner is down’

A significant outage has struck ChatGPT, leaving many users unable to receive responses from the popular AI chatbot. Instead of generating answers, the service failed to react to prompts, causing widespread frustration, particularly during the busy morning work period.

Owner OpenAI has officially launched an investigation into the mysterious malfunction of ChatGPT after its status page confirmed a problem was detected.

Over a thousand complaints were registered on the outage tracking site Down Detector. Social media was flooded with reports from affected users, with one calling it an unprecedented event and another joking that their ‘work partner is down’.

Instead of a full global blackout, initial tests suggested the issue might be limited to some users, as functionality remained for others.

If you find ChatGPT is unresponsive, you can attempt several fixes instead of simply waiting. First, verify the outage is on your end by checking OpenAI’s official status page or Down Detector instead of assuming your connection is at fault.

If the service is operational, try switching to a different browser or an incognito window to rule out local cache issues. Alternatively, use the official ChatGPT mobile app to access it.

For a more thorough solution, clear your browser’s cache and cookies, or as a last resort, consider using an alternative AI service like Microsoft Copilot or Google Gemini to continue your work without interruption.

OpenAI is working to resolve the problem. The company advises users to check its official service status page for updates, rather than relying solely on social media reports.

The incident highlights the growing dependence on AI tools for daily tasks and the disruption caused when such a centralised service experiences technical difficulties.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

OpenAI eyes India for large-scale AI infrastructure

According to Bloomberg, OpenAI is weighing partnerships in India to build a data centre of at least 1 gigawatt capacity as part of its Stargate project. Such a facility would represent one of Asia’s most significant AI infrastructure investments.

The company recently registered as a legal entity in India and is recruiting a local team. It also announced plans in August to open its first office in New Delhi later this year, highlighting the importance of India’s second-largest market by user base.

The prospective data centre is linked to Stargate, a private-sector AI investment programme valued at up to $500 billion and backed by SoftBank, OpenAI and Oracle. The project was first introduced in January by US President Donald Trump.

Details on the timing and location of the Indian facility remain unclear. Reports suggest that OpenAI chief executive Sam Altman could provide further information during a visit to India in September.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

ChatGPT safety checks may trigger police action

OpenAI has confirmed that ChatGPT conversations signalling a risk of serious harm to others can be reviewed by human moderators and may even reach the police.

The company explained these measures in a blog post, stressing that its system is designed to balance user privacy with public safety.

The safeguards treat self-harm differently from threats to others. When a user expresses suicidal intent, ChatGPT directs them to professional resources instead of contacting law enforcement.

By contrast, conversations showing intent to harm someone else are escalated to trained moderators, and if they identify an imminent risk, OpenAI may alert authorities and suspend accounts.

The company admitted its safety measures work better in short conversations than in lengthy or repeated ones, where safeguards can weaken.

OpenAI is working to strengthen consistency across interactions and developing parental controls, new interventions for risky behaviour, and potential connections to professional help before crises worsen.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!