AI news, Tuesday 27 October 2026

Guidewire agrees to buy London insurance AI firm hyperexponential for $1.7bn plus up to $400m earn-out

Insurance software maker Guidewire said it will acquire London-based hyperexponential for $1.7bn in cash, with up to $400m more tied to recurring revenue targets. Hyperexponential sells AI-native underwriting and pricing software to commercial property and casualty insurers. Guidewire expects completion in the second quarter of its fiscal 2027 and plans to part-fund the deal with a $1bn term loan.

UK angle: hyperexponential is a London-based company whose listed customers include Aviva and Beazley.

Washington Post: AI companies borrowing heavily and unsure where future funding will come from

The Washington Post reports that AI companies have been borrowing hundreds of billions of dollars, selling fresh stock and devising novel, potentially risky financing methods. The piece frames the sector as desperate for cash and unsure where the money will come from.

Privacy group noyb files five complaints against Scale AI over alleged worker surveillance

European privacy advocacy group noyb has lodged five complaints against Scale AI with regulators in France, Italy, the Netherlands and Austria. The complaints concern alleged surveillance of workers on its Outlier platform, who produce AI training data.

IBM launches identity and control tools for AI agents, with Vault 2.2 and Agent Identity now generally available

At its TechXchange 2026 event, IBM announced a set of tools for governing AI agents in the enterprise. Vault 2.2 and Agent Identity are generally available, giving agents verified identities and governed access. IBM says watsonx Orchestrate can discover agents across Amazon Bedrock, Microsoft Foundry and Google Gemini and add sandbox containment with a kill switch, though those discovery and containment features are only in tech preview.

Australian Signals Directorate warns AI is an 'unprecedented' cyber threat as warnings rise 80%

Australia's signals intelligence agency says in its annual cyber threat report that AI helps malicious actors find and exploit vulnerabilities faster and at greater scale. It reports an 80% rise in cyber security warnings over the past year, with more than 4,400 alerts issued. State-linked actors from China, Russia and North Korea are said to have targeted Australian organisations, which is a useful signal for defenders elsewhere.