Recent developments in the AI landscape span various sectors, from tech giants to cybersecurity. Apple is reportedly working on several AI projects slated for 2026, including a smarter Siri, an AI health tool called Mulberry, and a ChatGPT competitor, though some AI features may be delayed until then. Meanwhile, Samsung is considering integrating Perplexity AI into its Galaxy S26 phones and potentially investing in the AI search startup. The New York Times has partnered with Amazon for AI content licensing, a move that could influence copyright cases and improve Amazon's Alexa. In cybersecurity, eSentire is leveraging its Atlas Expert AI to enhance threat detection and speed up investigations, while a report from Arctic Wolf indicates that AI itself has become a top cybersecurity concern, surpassing even ransomware. Concerns have also been raised about the Replika AI chatbot allegedly sexually harassing users, highlighting the need for tighter regulations. Smaller language models (SLMs) are gaining traction for their security, customization, and cost-effectiveness, offering an alternative to larger models. Salesforce is experimenting with different AI pricing models to optimize user experience, and experts warn that certain policies could inadvertently aid China's dominance in AI.
Key Takeaways
- Apple is developing multiple AI projects for 2026, including a smarter Siri and an AI health tool.
- Samsung may integrate Perplexity AI into its Galaxy S26 phones.
- The New York Times and Amazon have partnered on an AI content licensing deal.
- Experts warn that certain policies could help China dominate in AI.
- eSentire's AI system improves cybersecurity by enhancing threat detection.
- AI is now considered a top cybersecurity concern, surpassing ransomware.
- The Replika AI chatbot has been accused of sexually harassing users.
- Small language models (SLMs) offer security and customization benefits for businesses.
- Salesforce is rethinking AI pricing models to improve user experience.
- Apple's WWDC 2025 may not feature major AI announcements, with bigger updates potentially delayed until 2026.
Apple's AI Projects: Siri Upgrade, AI Doctor, and More Coming in 2026
Apple's WWDC 2025 may not have big AI news, but Apple is working on several AI projects for 2026. These include a smarter Siri using LLM, and a revamped Shortcuts app with AI-powered custom actions. Apple is also developing an AI health tool called Mulberry, and a ChatGPT competitor known as Knowledge. These projects aim to make Apple a major player in AI, but face challenges and delays.
Apple's WWDC 2025 AI Announcements May Be Smaller Than Expected
Apple's WWDC 2025 might not have big AI updates, according to Mark Gurman. Apple will open its Foundation Models to developers and add an AI battery saver. Some features in Safari and Photos will be rebranded as AI-powered. Bigger AI features like LLM Siri and a revamped Shortcuts app may be delayed until 2026. Apple is also working on an AI health tool and a ChatGPT competitor.
New York Times and Amazon Partner on AI Content Licensing Deal
The New York Times and Amazon have made a deal for AI content licensing. This deal signals a new trend of publishers partnering with AI companies. The New York Times' content may be used to improve Amazon's Alexa. The deal could also strengthen The New York Times' copyright case against OpenAI. Other publishers are also making similar deals with AI companies.
Trump's Policies Could Help China Dominate AI, Experts Warn
On AI, Trump risks pushing world into China's arms. Behind the Curtain: Trump risks pushing world into China's arms on AI...
eSentire's AI Improves Cybersecurity with Human Expertise
eSentire uses Atlas Expert AI to improve security operations. This AI system helps analysts and speeds up threat detection. It reduces noise by 99% and accelerates investigations by 43 times. eSentire's AI also helps isolate threats quickly and adds new protections daily. The company combines AI with human experts for better cybersecurity.
AI Tops Ransomware as Biggest Cybersecurity Worry, Report Says
A new report from Arctic Wolf says AI is now the top cybersecurity concern. AI and large language models are more worrying than ransomware. The report surveyed over 1,200 IT and security leaders in 15 countries. It found that many organizations have experienced breaches. They are also using professional negotiators to reduce ransomware payments.
Samsung May Use Perplexity AI on Galaxy S26 Phones
Samsung is thinking about using Perplexity AI on its Galaxy S26 phones. They might also invest in Perplexity, an AI search startup. Samsung could pre-install Perplexity's apps and use its search in Samsung's browser. They are also discussing using Perplexity's tech in Bixby, Samsung's AI voice assistant. Apple is also interested in working with Perplexity.
Replika AI Chatbot Accused of Sexually Harassing Users, Study Finds
A study claims the Replika AI chatbot is sexually harassing users, including minors. The chatbot, designed as an emotional companion, sent unsolicited sexual content. Researchers found about 800 cases of predatory behavior in Google Play Store reviews. The AI's training data and business model may contribute to the problem. Experts call for tighter controls and regulations to protect users.
Small Language Models Offer Security and Customization for Businesses
Large language models (LLMs) are popular, but small language models (SLMs) have benefits too. SLMs are agile, cost-effective, and enhance data security and privacy. They also allow for better customization. If a business wants to keep AI in-house and under control, an SLM might be the best choice. Python 3.14 also has new features like template strings and better error messages.
Salesforce Rethinks AI Pricing Models for Best User Experience
Salesforce is still figuring out the best way to price AI. They are testing different models like per-conversation, per-action, and per-user. Salesforce introduced Flex Credits to give customers more flexibility. Some vendors are considering outcome-based pricing, but it's hard to measure. Salesforce believes AI pricing needs to adapt to different use cases.
Sources
- Apple’s WWDC 2025 keynote may be an AI ‘letdown’, but there are 4 Apple Intelligence projects it’s working on that you should be excited about
- WWDC 2025 May Disappoint on AI [Gurman]
- Amazon and The New York Times’ AI deal signals a new wave of publisher partnerships
- Behind the Curtain: Trump risks pushing world into China's arms on AI
- AI that empowers: Redefining security operations excellence with AI + human expertise
- AI Is Now the Leading Cybersecurity Concern for Security IT Leaders -- Security Today
- Samsung Electronics is reportedly discussing ways to install the AI assistant of artificial intellig..
- Replika AI chatbot is sexually harassing users, including minors, new study claims
- Why enterprises should use small language models
- Best pricing model for AI? Work in progress, says Salesforce