Gartner warns that more than 40 percent of agentic AI projects could be cancelled by 2027

More than 40% of agentic AI projects will likely be cancelled by the end of 2027 due to rising costs, limited business value, and poor risk control, according to research firm Gartner.

These cancellations are expected as many early-stage initiatives remain trapped in hype, often misapplied and far from ready for real-world deployment.

Gartner analyst Anushree Verma warned that most agentic AI efforts are still at the proof-of-concept stage. Instead of focusing on scalable production, many companies have been distracted by experimental use cases, underestimating the cost and complexity of full-scale implementation.

A recent poll by Gartner found that only 19% of organisations had made significant investments in agentic AI, while 31% were undecided or waiting.

Much of the current hype is fuelled by vendors engaging in ‘agent washing’ — marketing existing tools like chatbots or RPA under a new agentic label without offering true agentic capabilities.

Out of thousands of vendors, Gartner believes only around 130 offer legitimate agentic solutions. Verma noted that most agentic models today lack the intelligence to deliver strong returns or follow complex instructions independently.

Still, agentic AI holds long-term promise. Gartner expects 15% of daily workplace decisions to be handled autonomously by 2028, up from zero in 2024. Moreover, one-third of enterprise applications will include agentic capabilities by then.

However, to succeed, organisations must reimagine workflows from the ground up, focusing on enterprise-wide productivity instead of isolated task automation.

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

Efforts to address internet fragmentation take centre stage at IGF 2025 in Norway

On the final day of the Internet Governance Forum 2025 in Lillestrøm, Norway, stakeholders from governments, civil society, technical communities, and the private sector gathered to launch the new work cycle of the Policy Network on Internet Fragmentation (PNIF). Now entering its third year, the PNIF unveiled a structured framework to analyse internet fragmentation across three dimensions: user experience, internet governance coordination, and the technical infrastructure layer.

The session emphasised the urgent need for international cooperation to counter growing fragmentation threats, as enshrined in paragraph 29C of the Global Digital Compact. Speakers raised alarm over how political and economic forces are re-shaping the global internet.

With internet shutdowns and digital censorship increasingly normalised as tools of state control—highlighted by Iran’s recent 90-million-person shutdown—concerns about sovereignty overriding openness were prominent. Michel Lambert described this shift as a ‘political normalisation of network control.’

Marilia Maciel, Director of Digital Trade and Economic Security at Diplo, emphasised how trade and investment policies fuel economic fragmentation. Cuts to internet freedom funding were highlighted by both Lambert and Joyce Chen, who noted severe consequences for underserved regions like the Pacific.

Marilia Maciel
Marilia Maciel, Director of Digital Trade and Economic Security at Diplo

From the technical community, Dhruv, representing the Internet Architecture Board, stressed the importance of safeguarding the internet’s interoperability by including technical experts in regulatory processes. Joyce Chen also pointed to successful coordination initiatives such as the Technical Community Coalition on Multi-Stakeholderism (TCCM).

Naim Gjokaj, State Secretary in Montenegro, offered a government perspective, advocating for stronger legal frameworks and regional coordination to avoid inadvertent fragmentation while supporting connectivity in rural areas.

The session concluded with a call to action: PNIF will focus its upcoming work on developing concrete, risk-based recommendations to implement the Global Digital Compact. Co-facilitators Sheetal Kumar and Bruna Santos encouraged broad community participation, aiming to deliver a final report by 1 November.

Despite the challenges, the atmosphere remained collaborative and forward-looking, reinforcing the importance of inclusive dialogue to ensure the internet remains a unified, accessible, and resilient resource for all.

Track all key moments from the Internet Governance Forum 2025 on our dedicated IGF page.

How can technical standards bridge or broaden the digital divide?

At the Internet Governance Forum 2025 in Lillestrøm, Norway, the Freedom Online Coalition convened a diverse panel to explore how technical standards shape global connectivity and inclusion. The session, moderated by Laura O’Brien, Senior International Counsel at Access Now, highlighted how open and interoperable standards can empower underserved communities.

Divine Agbeti, Director General of the Cybersecurity Authority of Ghana, shared how mobile money systems helped bring over 80% of Ghana’s adult population into the digital financial fold—an example of how shared standards translate into real-world impact, especially across Africa. However, the conversation quickly turned to the systemic barriers that exclude many from the standard-setting process itself.

ICANN’s At-Large Advisory Committee member emphasised challenges like high membership fees, lack of transparency, English-only proceedings, and complex technical jargon.

Stephanie Borg Psaila, Director of Digital Policy at Diplo, presented detailed research mapping these hurdles across bodies like ITU, ICANN, and IETF, and called for reforms such as multilingual interpretation, hybrid meeting formats, and adjusted membership models to enable broader civil society participation.

Stephanie Borg Psaila

Security and infrastructure governance also featured prominently. Rose Payne, Policy and Advocacy Lead at Global Partners Digital, spotlighted the role of technical standards in safeguarding subsea cables—which carry 95–99% of transnational data—but also pointed to the limitations of technical solutions when facing geopolitical threats.

She underscored the urgency of updating international legal frameworks like UNCLOS and fostering cooperation between governments, the private sector, and civil society. Alex Walden, Global Head of Human Rights at Google, also reaffirmed the private sector’s role in investing in global connectivity while advocating for human rights-based frameworks and inclusive multistakeholder participation.

While the session closed on a constructive note, tensions emerged during the Q&A. Technical community members like Colin Perkins (University of Glasgow) and Harold, a technologist and civil society advocate, challenged the panel’s framing, arguing that distinctions between technical and civil society actors are often artificial and counterproductive.

Panellists responded diplomatically, acknowledging the need for more nuanced engagement and mutual understanding. Despite differing views, the forum concluded with shared commitments: dismantling barriers to participation, building cross-sectoral capacity, and grounding technical governance in international human rights from the outset.

Track all key moments from the Internet Governance Forum 2025 on our dedicated IGF page.

YouTube adds AI search results for travel, shopping and more

YouTube is launching a new AI-powered search feature that mirrors Google’s AI Overviews, aiming to improve how users discover content on the platform.

The update introduces an ‘AI-powered search results carousel’ when YouTube users search for shopping, travel, or local activities.

The carousel offers a collection of video thumbnails and an AI-generated summary highlighting the key topics related to the search. For example, someone searching for ‘best beaches in Hawaii’ might see curated clips of snorkelling locations, volcanic coastlines, and planning tips — all surfaced by the AI.

Currently, the feature is available only to YouTube Premium users in the US. However, the platform plans to expand its conversational AI tool — which provides deeper insights, suggestions, and video summaries — to non-Premium users in the US soon.

That tool was first launched in 2023 to help users better understand content while watching.

YouTube is doubling down on AI features to keep users engaged and make content discovery more intuitive, especially in categories involving planning and decision-making.

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

DeepSeek struggles to launch R2 amid NVIDIA chip shortage

The launch of DeepSeek’s next-generation AI model, R2, is expected to face delays due to a shortage of NVIDIA H20 chips in China.

These chips, designed specifically for the Chinese market following US export restrictions, are essential for running DeepSeek’s highly optimised models.

The ban on H20 shipments in April has triggered widespread concern among cloud providers about the scalability of R2, especially if it outperforms existing open-source models.

CEO Liang Wenfeng has reportedly held back the model’s release, expressing dissatisfaction with its current performance.

Engineers continue refining R2, but the lack of compatible hardware poses a deeper challenge. DeepSeek’s reliance on NVIDIA architecture makes switching to Chinese chips inefficient, as the models are tightly built for NVIDIA’s software and hardware ecosystem.

Some Chinese firms have begun using workarounds by flying engineers to Malaysia, where NVIDIA chips are still available in local data centres.

After training their models abroad, teams return to China with trained systems. Others rely on gaming GPUs like the RTX 5090, which are easier to access via grey markets despite restrictions.

While Chinese tech giants ordered 1.2 million H20 chips earlier in 2025 to meet demand sparked by R1’s success, inventory is still unlikely to support a full R2 rollout.

Companies outside China may launch R2 more easily without facing the same export hurdles.

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

Hawaiian Airlines confirms flights are safe despite cyberattack

Hawaiian Airlines has reported a cyberattack that affected parts of its IT infrastructure, though the carrier confirmed all flights remain unaffected and are operating as scheduled.

Now part of the Alaska Air Group, the airline stated it is actively working with authorities and cybersecurity experts to investigate and resolve the incident.

In a statement, the airline stressed that the safety and security of passengers and staff remain its highest priority. It has taken steps to protect its systems, restoring affected services while continuing full operations. No disruption to passenger travel has been reported.

The exact nature of the attack has not been disclosed, and no group has claimed responsibility so far. The Federal Aviation Administration (FAA) confirmed it monitors the situation closely and remains in contact with the airline. It added that there has been no impact on flight safety.

Cyberattacks in aviation are becoming increasingly common due to the sector’s heavy reliance on complex digital systems. Earlier incidents this year included cyberattacks on WestJet and Japan Airlines, which caused operational disruptions but did not compromise passenger data.

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

Le Chat leads AI privacy ranking report

A new report has revealed that Le Chat from Mistral AI is the most privacy-respecting generative AI, with ChatGPT and Grok close behind. The study by Incogni assessed nine popular services against 11 criteria covering data use, sharing and transparency.

Meta AI came last, flagged for poor privacy practices and extensive data sharing. According to the findings, Gemini and Copilot also performed poorly in protecting user privacy.

Incogni highlighted that several services, including ChatGPT and Grok, allow users to stop their data from being used for training. However, other providers like Meta AI, Pi AI and Gemini offered no clear way to opt-out.

The report warned that AI firms often share data with service providers, affiliates, researchers and law enforcement. Clear, readable privacy policies and opt-out tools were key for building trust.

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

BT report shows rise in cyber attacks on UK small firms

A BT report has found that 42% of small businesses in the UK suffered a cyberattack in the past year. The study also revealed that 67% of medium-sized firms were targeted, while many lacked basic security measures or staff training.

Phishing was named the most common threat, hitting 85% of businesses in the UK, and ransomware incidents have more than doubled. BT’s new training programme aims to help SMEs take practical steps to reduce risks, covering topics like AI threats, account takeovers and QR code scams.

Tris Morgan from BT highlighted that SMEs face serious risks from cyber attacks, which could threaten their survival. He stressed that security is a necessary foundation and can be achieved without vast resources.

The report follows wider warnings on AI-enabled cyber threats, with other studies showing that few firms feel prepared for these risks. BT’s training is part of its mission to help businesses grow confidently despite digital dangers.

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

IGF and WSIS platforms must be strengthened, not replaced, say leaders

At the Internet Governance Forum 2025 in Lillestrøm, Norway, stakeholders gathered to assess the International Telecommunication Union’s (ITU) role in the WSIS Plus 20 review process.

The session, moderated by Cynthia Lesufi of South Africa, invited input on the achievements and future direction of the World Summit on the Information Society (WSIS), now marking its 20th year.

Speakers from Brazil, Australia, Korea, Germany, Japan, Cuba, South Africa, Ghana, Nigeria, and Bangladesh offered their national and regional insights.

There was strong consensus on maintaining and strengthening existing platforms like the Internet Governance Forum (IGF) and WSIS Forum, rather than creating new mechanisms that might burden developing countries.

Renata Santoyo, representing Brazil’s telecommunications regulator ANATEL, affirmed ITU’s coordinating role: ‘The WSIS architecture remains valuable, and ITU has been instrumental in supporting its action lines.’

Australia’s William Lee echoed this, commending ITU’s work on integrating WSIS with the SDGs and the Global Digital Compact, and noted: ‘The digital divide is now less about access and more about usability.’

Korean vice chair of the ITU Council Working Group, Mina Seonmin Jun, stressed the continued inequality in her region: ‘One third of Asia-Pacific remains offline. WSIS must go beyond infrastructure and focus on equity.’

 People, Person, Cinema, Adult, Female, Woman, Male, Man, Crowd, Face, Head, Indoors, Audience, Chair, Furniture

Swantje Jager Lindemann from Germany backed extending the IGF mandate without renegotiation, saying: ‘The mandate is broad enough. What we need is better support and sustainable funding.’

Japan’s Yoichi Iida, former vice minister and now special advisor, also warned against reopening existing mandates, instead calling for a stronger IGF secretariat. ‘We must focus on inclusivity, not duplicating structures,’ he said.

ITU’s Gitanjali Sah outlined its leadership on WSIS action lines, noting the organisation’s collaboration with over 50 UN bodies. ‘2.6 billion people are still offline. Connectivity must be meaningful and inclusive,’ she said, highlighting ITU’s technical support on cybersecurity, capacity building, and standards.

Cuba’s representative stressed that the WSIS outcome documents remain fully valid and should be reaffirmed rather than rewritten. ‘Creating new mechanisms risks excluding countries with limited resources,’ they argued.

Local voices called for grassroots inclusion. Louvo Gray from the South African IGF asked, ‘How do we ensure marginalised voices from the Global South are truly heard?’ Ghana’s Kweku Enchi proposed tapping retired language teachers to bridge digital and generational divides.

Abdul Karim from Nigeria raised concerns about public access to the review documents. Sah confirmed that most contributions are published on the ITU website unless requested otherwise.

The UNDP representative reiterated UN-wide support for an inclusive WSIS review, while Mohamed Abdulla Konu of Bangladesh IGF pressed for developing countries’ voices to be meaningfully reflected.

Speakers agreed that the WSIS Plus 20 review is a key opportunity to refocus digital governance on inclusion, equity, and sustainability. The ITU will submit the compiled inputs to the UN General Assembly in December, while South Africa will include the session’s outcomes in its high-level report.

Track all key moments from the Internet Governance Forum 2025 on our dedicated IGF page.

Tower of Babel reimagined: IGF 2025 experiment highlights language barriers in internet governance

At the 2025 Internet Governance Forum in Lillestrøm, Norway, an unconventional session titled ‘Tower of Babel Chaos’ challenged the norm of using English as the default language in global digital policy discussions. Moderator Virginia Paque, Senior Policy Editor of Diplo and the only native English speaker among the participants, suspended English as the session’s required language and encouraged attendees to define internet governance and interact in their own native tongues.

That move sparked both confusion and revelation as participants experienced firsthand the communicative fragmentation caused by linguistic diversity. The experiment led to the spontaneous clustering of speakers into language groups and highlighted the isolation of individuals whose languages—such as Maltese, Samoan, Cape Verdean Creole, and Chichewa—had no other representation.

Participants reported feelings ranging from curiosity to frustration, underlining the practical importance of shared language in international settings. Yet, some also discovered unexpected bridges through linguistic overlap or body language, hinting at the potential for cross-cultural communication even in chaotic conditions.

 Stage, Indoors, Theater, Architecture, Building, Cinema, Lighting, Auditorium, Hall, Person, Electronics, Screen, People

AI emerged as a potential remedy. Ken Huang from Lingo AI noted that while AI can process thousands of languages, its effectiveness is currently limited by a lack of diverse datasets, making it default to English and other dominant tongues. Others emphasised that while technology offers hope—like real-time translation tools—it cannot guarantee equitable inclusion for all linguistic groups, particularly under-resourced languages.

The session ultimately balanced idealism with pragmatism. While many acknowledged the convenience of English as a global lingua franca, others argued for providing multiple language options with simultaneous interpretation, as practised by institutions like the UN.

The discussion underscored the political, cultural, and technological complexities of multilingualism in internet governance, and concluded with a shared recognition: fostering a more inclusive digital dialogue means embracing both innovation and linguistic diversity.

Track all key moments from the Internet Governance Forum 2025 on our dedicated IGF page.