A new global cybersecurity report by Semperis reveals that 84% of ransomware victims experienced attacks where criminals used AI. Surveying 1,200 IT and security professionals across nine countries, the study found that AI helps attackers automate tasks, craft convincing phishing messages and find security weaknesses. More than 69% of victims faced multiple attacks, and 59% paid a ransom at least once. Former national cyber director Chris Inglis warned that autonomous ransomware attacks are already happening.
On the consumer side, Accenture's annual Holiday Shopping Survey found that 76% of consumers would use an AI agent as a shopping partner. Nearly 3 in 10 trust an AI agent more than their own choices, and consumers estimate GenAI will influence 45% of their holiday spending. However, only 24% would let an AI agent decide what to buy, showing people still prefer keeping control over their purchases.
Numerator's AI Consumer Index identified health and beauty, apparel, footwear, accessories and electronics as the product categories most disrupted by AI shopping. AI-engaged households account for $37 billion in additional annual spending, and over 30% of U.S. households now use AI weekly. Costco, TJX Companies and Amazon over-index among AI-engaged consumers, while Chipotle, Panera Bread and Jack in the Box lead in the restaurant space.
OpenAI accused a former employee of stealing intellectual property related to an unreleased device and sharing that information with Apple after leaving the company. OpenAI did not name the employee or describe their role, and the company has not said whether it will take legal action. The case raises questions about intellectual property protections in the competitive AI hardware space.
Australia's Industry and Science Minister Tim Ayres announced support for hosting AI training and building models locally under Australian laws. The government plans to launch a A$50 million AI Accelerator under the Cooperative Research Centre program in early 2027. The minister said industrial and research policies will help spread AI adoption across businesses while strengthening national security and technical expertise.
In academic and product news, the Smithsonian Astrophysical Observatory received $316,000 as part of a $988,000 AI for Math Fund award from Renaissance Philanthropy and XTX Markets. The funding will support new AI tools for analyzing large astronomical datasets. Separately, Abacus AI launched products including ChatLLM, which gives access to models like GPT-5.5, Sonnet 4.6 and Gemini 3.1 Pro for $10 per month, and an AI Agent for multi-step tasks with Gmail, Slack and Google Workspace at $20 per month.
Key Takeaways
- <ul><li>84% of ransomware victims experienced AI-assisted attacks, according to a Semperis report surveying 1,200 IT professionals across nine countries</li><li>76% of consumers would use an AI agent for holiday shopping, but only 24% would let it make purchase decisions independently</li><li>Health and beauty, apparel, electronics and accessories face the greatest AI shopping disruption, with AI-engaged households driving $37 billion in extra annual spending</li><li>OpenAI accused a former employee of sharing unreleased device secrets with Apple; no legal action has been announced</li><li>Australia plans a A$50 million AI Accelerator program launching in early 2027 to support local AI training and model development</li><li>Smithsonian Astrophysical Observatory received $316,000 from a $988,000 AI for Math Fund award for analyzing astronomical datasets</li><li>Over 69% of ransomware victims experienced multiple attacks and 59% paid a ransom at least once</li><li>Abacus AI offers ChatLLM at $10/month and an AI Agent at $20/month with access to GPT-5.5, Sonnet 4.6 and Gemini 3.1 Pro</li><li>40% of ransomware attacks involved physical and psychological threats against victims, and 44% exploited victims' own AI systems</li><li>Costco, TJX Companies and Amazon over-index among AI-engaged consumers, while Chipotle, Panera Bread and Jack in the Box lead in restaurant AI adoption</li></ul>
AI used in 84% of ransomware attacks, global report reveals
A new global cybersecurity report by Semperis found that 84% of ransomware victims experienced attacks where AI was used by criminals. The report surveyed 1,200 IT and security professionals across nine countries. More than a third of victims also faced psychological or physical threats against employees. Former cyber director Chris Inglis warned that autonomous ransomware attacks are already happening.
Study finds AI involved in over 84% of ransomware attacks
A Semperis report shows that cybercriminals are increasingly using AI to make ransomware attacks more targeted and harder to detect. AI helps attackers automate tasks, craft convincing phishing messages and find security weaknesses. More than 69% of ransomware victims experienced multiple attacks, and 59% paid a ransom at least once. The report stresses the need for stronger cybersecurity strategies in businesses.
84% of ransomware victims say AI was used in attacks
A new survey found that over 80% of ransomware victims said attackers weaponized AI in their schemes. Chris Inglis, former national cyber director, said AI makes attacks faster and more sophisticated. About 44% of victims experienced attacks that exploited their own AI systems. The report also revealed that 40% of attacks involved physical and psychological threats against victims.
Accenture: 76% of consumers open to AI agent for holiday shopping
Accenture's annual Holiday Shopping Survey found that 76% of consumers would use an AI agent as a shopping partner. Nearly 3 in 10 consumers trust an AI agent more than their own choices. Consumers estimate GenAI will influence 45% of their holiday spending. However, only 24% would let an AI agent decide what to buy, showing people prefer keeping control.
Numerator reveals product categories most disrupted by AI shopping
Numerator's AI Consumer Index found that health and beauty, apparel, footwear, accessories and electronics face the greatest AI disruption. AI-engaged households account for $37 billion in additional annual spending. Over 30% of U.S. households now use AI weekly. Costco, TJX Companies and Amazon over-index among AI-engaged consumers, while Chipotle, Panera Bread and Jack in the Box lead in restaurants.
LLM agents share client goals through KV cache
Researchers tested OLMo-3-7B-Instruct in a two-client setting with 20 market scenarios. The model deviated from the current client's objective in 22% of eligible decisions when sharing conversational memory. Linear probes found client-goal ownership in middle KV cache layers and specific attention heads. Ablating competitor goal representations shifted decisions toward the current client. This highlights the need for relationship-specific state isolation in persistent LLM agents.
AirTrunk and Firmus stumble amid AI boom doubts
AirTrunk, a key player in the AI boom, faced a major setback that raised questions about the sector's sustainability. Rival Firmus also encountered challenges in replicating AirTrunk's success. The AI industry has seen several high-profile startup failures despite heavy global investment. Many investors still believe in AI's potential returns, but recent setbacks call for a more careful look at the risks involved.
100 free prompts to fix bad AI-generated flyers
An article offers 100 free prompts to help AI produce better flyers and posters. The author was inspired by a bland Halloween school flyer that felt like every other one. Bad AI design is becoming common, and some businesses are facing customer frustration. The prompts aim to help users break away from repetitive and unappealing AI-generated visuals.
Smithsonian Astrophysical Observatory wins AI for Math award
Cecilia Garraffo, a CfA astrophysicist and AstroAI director, received $316,000 as part of a $988,000 AI for Math Fund award. The funding comes from Renaissance Philanthropy and founding donor XTX Markets. The award will support new AI tools for analyzing large astronomical datasets. The Smithsonian Astrophysical Observatory has over 80 years of research history.
Abacus AI platform review and practical use cases
Abacus AI offers products including ChatLLM, Abacus AI Agent, and Personal AI Agents. ChatLLM gives access to models like GPT-5.5, Sonnet 4.6, and Gemini 3.1 Pro for $10 per month. The AI Agent handles multi-step tasks with Gmail, Slack, and Google Workspace, with Pro at $20 per month. Use cases include document analysis, website development, CRM, trading, and language learning.
UMD Smith School launches AI-powered marketing master's program
The University of Maryland's Robert H. Smith School of Business will launch a one-year marketing master's program in fall 2027. The 30-credit program combines artificial intelligence with marketing fundamentals and hands-on projects. Students will study brand strategy, consumer insights and client management with an emphasis on AI applications. Graduates can pursue careers in brand development, marketing strategy and campaign management.
OpenAI accuses ex-employee of sharing AI device secrets with Apple
OpenAI accused a former employee of stealing intellectual property related to an unreleased device. The company said the ex-worker shared some of that information with Apple after leaving OpenAI. OpenAI did not name the employee or describe their role. The company has not said whether it will take legal action.
Australia backs local AI training with new funding program
Australia's Industry and Science Minister Tim Ayres supports hosting AI training and building models locally under Australian laws. He said this approach would strengthen innovation, national security and technical expertise. The government plans to launch a A$50 million AI Accelerator under the Cooperative Research Centre program in early 2027. The minister added that industrial and research policies will help spread AI adoption across businesses.
Sources
- AI linked to 84% of ransomware attacks, new report finds
- Study finds AI involved in more than 84% of Ransomware Attacks
- 84 percent of ransomware victims point to AI link: Report
- Accenture: 76% of consumers are open to working with an AI agent for holiday shopping
- Numerator: The product categories most likely to be disrupted by AI are…
- When Goals Leak Across Relationships: How Cached Goal Representations Shape Cross-Client Decisions
- AirTrunk, Firmus stumbles test investors’ faith in the AI boom amid market volatility and rising debt costs
- Your AI-Generated Flyers Look Terrible. Here Are 100 Prompts to Fix Them in Seconds
- Smithsonian Astrophysical Observatory Wins AI For Math Fund Award | Center for Astrophysics
- Abacus AI in Action Review: Major Products & Practical Use Cases
- UMD Smith School Introduces AI-Driven MS in Marketing Program for Fall 2027
- OpenAI accuses former employee of leaking intel on AI device to Apple
- Australian minister backs local AI training to build national capability | MLex | Specialist news and analysis on legal risk and regulation
Comments
Please log in to post a comment.