The Distillery: Hackers Using AI for Cyberattacks and Hong Kong Telecom Exits US Market

November 2026

Identity & Security

Juniper Research

Analyst House

In this edition:

• AI in cybercrime: Hackers are now using AI to leverage smart devices and carry out sophisticated cyberattacks
  
• Free report: The state of chargeback management - how AI is reshaping performance
  
• Hong Kong Telecom exits US market: Regulatory pressures are dubbed as the primary reason behind their departure
  
• Mastercard partners with OpenAI: The joint venture will likely play a key role in Mastercard's agentic commerce offering
  
• Free report: Stopping scams at the source - the case for a whole-of-ecosystem approach

IDENTITY & SECURITY

Hackers Using AI Are Attacking Smart Fridges, Exchanges, and Critical Infrastructure

On 29th October 2026, Binance, the world’s largest cryptocurrency exchange, suffered its largest hack, when attackers stole approximately $6.3 billion in digital assets across several blockchains. This is the largest cryptocurrency theft on record, with the breach being traced to a tampered installer for an AI coding tool downloaded by a Binance engineer. Only four days later, Brazil’s largest cryptocurrency exchange, Mercado Bitcoin, suffered a security breach which led to the loss of nearly $40 million.

On 21st October 2026, researchers surfaced two critical vulnerabilities affecting PHP, a popular server-side programming language powering approximately 70% of the world’s websites, alongside the patches to fix these. Due to the severity and scale of these vulnerabilities, many servers, sites and pages were taken offline, or reduced to ‘limited’ functionality, as security personnel rushed to deploy the patches.

Separately, since early October 2026, hundreds of supermarkets across Europe have been hacked, with attackers pushing freezer temperatures past 50 degrees Celsius, and authorities investigating what they think is a large-scale cyberattack on food supply. Stores from Sweden, Belgium, France, and Ireland were reportedly compromised.

Whilst this handpicked list of cyber incidents may appear unrelated, they share a common, underlying theme: Artificial Intelligence being used by hackers to reduce the time needed to find and exploit such weaknesses

Distilled…

🟣 Hackers are leveraging the latest AI models to find and expose vulnerabilities in devices connected to the Internet, whether it is a smart refrigerator, or servers protecting billions in customer assets. AI agents or ‘swarms’ can run autonomously, around-the-clock, without getting tired. Furthermore, work is usually distributed across many agents, complete with their own internal communication and coordination mechanisms. This means the pace at which vulnerabilities in systems and devices are being exposed and exploited is far outpacing the ability to patch them.

🟣 AI can be used offensively, and defensively. The same techniques and models used in trying to find exploits and vulnerabilities are also used proactively to find and fix them before it is too late. Findings show that the PHP vulnerability was reported in August 2026, yet kept secret whilst researchers worked on a patch. It was only in October 2026, when Amazon’s Threat Intelligence detected an exploit attempt, that, several days later, an updated version of PHP was released. 

🟣 Prevention is better than cure. It is difficult for everyday users of technology to remain current with its daily exploits and vulnerabilities. However, small actions, such as regularly updating devices to the latest firmware, not clicking on unknown emails, links or attachments, and keeping highly sensitive or valuable data on devices not connected to the Internet, can protect users from avoidable harm.

READ ONLINE

MORE SECURITY INSIGHTS

🔐 AppSec Enterprise Spend to Approach $13 Billion Globally by 2031, as Market Leaders Revealed in New Competitor Leaderboard [Research]

🫆 Non-human Identities ~ Transforming Identity & Access Management [Whitepaper]

🪪 Digital Identity Verification in an Era of AI, Fraud & Regulatory Change [Whitepaper]

NEW REPORT

The State of Chargeback Management: How AI Is Reshaping Performance

Juniper Research recently partnered with Chargeflow to understand the impact of chargebacks on merchants and how enterprises in the industry use AI tools. The report discusses key findings, such as the value of chargebacks being raised will reach almost $57 billion annually by 2031.

The report also answers crucial questions, such as:

🟣 How are chargeback volumes changing over time? The change in the number or volume of chargeback requests over the years.

🟣 How are enterprises tackling chargebacks, and is this working? Is the management of chargebacks performed internally, or is it outsourced.

🟣 To what extent are enterprises using AI in chargeback management, and what results are they seeing? The impact of AI in chargeback management.

🟣 What do enterprises want from a solution? The ideal solution for an enterprise to address chargebacks.

🟣 Are enterprise merchants worried about the holiday season? Understanding whether the December holidays are particularly difficult, in terms of seeing an increase in chargebacks.

GET THE REPORT

TELECOMMUNICATIONS & CONNECTIVITY

Hong Kong Telecom to Exit US Market Following Regulatory Pressure

Hong Kong-based telco HKT has agreed to end its operations in the US, following pressure from the Federal Communications Commission (FCC), which has cited national security concerns over HKT’s links with Chinese operators.

HKT is a subsidiary of Hong Kong conglomerate PCCW, which is part-owned by China Unicom. The FCC has previously banned China Unicom, along with its rivals China Mobile and China Telecom, from providing international telco services to the US, citing concerns over fraud and national security.

Distilled…

🟣 The FCC is ramping up pressure on China’s telecoms companies, announcing plans to ban Chinese firms - as well as any operators using equipment from Chinese vendors (such as Huawei and ZTE) - from direct access to US phone numbers, on the grounds of national security. The regulator is also considering barring Chinese telco firms from connecting to US networks using Voice over Internet Protocol (VoIP), and from operating any data centres in the US.

🟣 The FCC claims the measures are to strengthen safeguards against spam calls, with Chair Brendan Carr saying it would bolster the ‘first line of defence against bad actors and illegal robocallers’. But the measures are also part of a wider crackdown on Chinese tech firms by the Trump administration, with the FCC recently voting to ban all Chinese labs from testing US electronics, beginning in December 2028.

🟣 China’s embassy in Washington criticised the proposal, saying it ‘…severely hinders normal economic and trade cooperation between Chinese and American businesses, and is not in the interest of any party, including US companies and consumers’. China has consistently denied any wrongdoing, and the escalation in restrictions is regarded by many as economic protectionism designed to support American and European vendors. Large parts of the EU and the UK have also banned Huawei and ZTE hardware from their networks, calling them ‘high-risk vendors’.

MORE TELECOMMUNICATIONS INSIGHTS

📶 Nokia and NVIDIA Edge AI-RAN Closer to Reality [Blog] 

💬 Rich Messaging, Real Growth: the Rise of RCS and OTT in CPaaS [Whitepaper]   

📡 Retail Roaming Revenue to Fall $4 Billion By 2031, as Regulatory Restrictions and Competition Bite [Research]

NEW REPORT

Stopping Scams at the Source: the Case for a Whole-of-ecosystem Approach

A collaborative report by Juniper Research, the Financial Technology Association (FTA), and Revolut found that Americans will see nearly 1.5 trillion scam ads and receive 28 billion fraudulent texts a year by 2031, without stronger action at the source. The report advocates that banks and fintechs cannot be the only ones defending against fraud. Stopping fraud requires a whole-of-ecosystem approach, which means better advertiser verification, faster removal of scam ads, stronger text message filtering, and easier ways for consumers to report fraud.

The research delves into:

🟣 Scam ads - These are deceptive, paid posts which mislead users into giving money, personal information, or account access.

🟣 Social media losses - Fake social media posts are driving losses for Americans.

🟣 Fraudulent texts - Scammers use mobile messages to draw in and defraud victims.

GET THE REPORT

 

FINTECH & PAYMENTS

Mastercard Partners with OpenAI

During its third-quarter Earnings meeting, Mastercard disclosed that it had entered an agreement with OpenAI. Mastercard’s CEO, Michael Miebach, said that this partnership ‘establishes collaboration across payments, joint services growth, and AI transformation’ and ‘goes well beyond’ Agent Pay. This announcement follows Mastercard’s trend of being an early adopter; building on its April 2025 launch of Agent Pay, and its first live authenticated agentic transaction in Singapore in March 2026. 

Even with Miebach emphasising that the scope of the partnership will extend beyond Agent Pay, this collaboration will likely play a key role in the expansion of Mastercard’s agentic commerce offering.

Distilled…

🟣 Building trust is the key to consumer adoption - Mastercard has framed agentic commerce adoption as requiring four conditions: trust in the payment, trust in the agent, trust in the ecosystem, and reach across merchants. Making up three of the four prerequisites, building trust will be an essential part of Mastercard’s strategy. Mastercard must use its experience in payments to provide protections for merchants and consumers, and utilise its reputation to secure consumer and merchant confidence.

🟣 Mastercard pulling ahead of competitors? - This partnership comes on the back of a series of moves to build Mastercard’s agentic commerce offering. These actions have led to Mastercard becoming one of the early leaders in the agentic commerce market. Evidence of this includes Mastercard performing its first live agentic transaction several months before rivals such as Visa, and being ranked as the leader in Juniper Research’s 2026 Agentic Commerce Payments Infrastructure Competitor Leaderboard.

🟣 AI has payment use cases beyond agentic commerce - It is noteworthy that Miebach has put a strong emphasis on this partnership going beyond agentic commerce, as there is a range of other payment use cases which would benefit from the integration of AI. Examples include the use of AI to monitor fraud activity, or AI-powered smart routing of payments. This broader range of features will be important for AI vendors, such as OpenAI, as it will enable them to offer services to a wider range of fintechs, beyond those which wish to offer agentic commerce.

GET THE WHITEPAPER

 

THE LAST DROP

🤖 Open AI Wants to Transform Banking: Why Are JPMorgan and Citi Pushing Back? [Blog]

🫆 Vibe Coding: The Fastest-growing Threat to Application Security Today [Whitepaper]

🏛️ Accounts Payable Automation ~ How Businesses Are Reducing Operational Costs [Whitepaper]

Subscribe

Contact Us