Several new AI tools are aiming to ease the burden on security and network operations teams. Netskope launched AgentSkope, a set of six AI agents that automate tasks like analyzing data loss prevention alerts and investigating insider threats. While these agents can create tickets and automate investigations, human analysts retain control over final responses. Separately, WatchGuard introduced Rai, an AI digital worker designed as an always-on security analyst for managed service providers (MSPs). Rai can monitor environments, correlate activity, and take pre-authorized actions in real time without human intervention, helping to reduce after-hours alert handling and free up analysts for higher-value work.
AI's growing role in the workplace and daily life is also raising new legal and ethical questions. A lawsuit called Doe v. Perplexity accuses Perplexity, along with Meta and Google, of sharing users' chatbot conversations with ad trackers without consent, even in Incognito Mode. In another case, five major publishers and author Scott Turow have sued Meta, alleging the company used copyrighted works to train its AI without permission, and that AI-generated books are flooding marketplaces like Amazon. Meanwhile, a report highlights how AI labels meant to identify fake content are failing, using the example of Emily Hart, a fake influencer created with Google Gemini who gained thousands of followers.
On the development side, Cursor introduced Composer autoinstall, a system that uses AI to automatically set up complex development environments for training other AI models. The system works in two stages: a goal-setting agent proposes commands and expected outputs, then a separate agent builds the environment. Composer 2 scored 61.7 percent on a benchmark test, up from 47.9 percent for the previous version. In a lighter moment, venture capitalist Marc Andreessen was mocked after a custom prompt he wrote for an AI chatbot was made public, in which he told the chatbot to never hallucinate—a request experts say misunderstands how large language models work.
In sales, Proshort uses AI to coach teams in real time by analyzing calls and providing specific feedback. Founder Gaurav Mishra says the tool has improved win rates by up to 20 percent and saves sales teams up to 10 hours per week on administrative tasks. Proshort has over 45 customers and crossed $1 million in annual recurring revenue within a year. Finally, a new report examines how AI is being used in health insurance prior authorization and claims review, highlighting risks like inaccurate or biased outcomes and privacy breaches, while also reviewing the Trump administration's AI framework that promotes development with limited federal restrictions.
Key Takeaways
- Netskope launched AgentSkope, six AI agents for SOC and NOC teams that automate alert analysis and insider threat investigations while keeping humans in control.
- WatchGuard introduced Rai, an AI digital worker for MSPs that monitors environments and takes pre-authorized actions in real time without human help.
- A lawsuit (Doe v. Perplexity) accuses Perplexity, Meta, and Google of sharing users' chatbot conversations with ad trackers without consent, even in Incognito Mode.
- Five major publishers and author Scott Turow sued Meta, alleging unauthorized use of copyrighted works to train AI and that AI-generated books are flooding Amazon.
- AI labels meant to identify fake content are failing, as shown by the fake influencer Emily Hart, created with Google Gemini, who gained thousands of followers.
- Cursor's Composer autoinstall uses AI to set up development environments for training other AI models; Composer 2 scored 61.7% on a benchmark, up from 47.9%.
- Marc Andreessen was mocked for telling an AI chatbot to never hallucinate, revealing a misunderstanding of how large language models work.
- Proshort, founded by Gaurav Mishra, uses AI to coach sales teams in real time, improving win rates by up to 20% and saving 10 hours per week; it crossed $1 million in ARR within a year.
- A report on AI in health insurance claims highlights risks of inaccurate or biased outcomes and privacy breaches, and reviews the Trump administration's AI framework with limited federal restrictions.
Netskope launches AgentSkope AI agents for SOC and NOC teams
Netskope introduced AgentSkope, a new set of AI agents designed to help security and network operations teams. The first release includes six agents that handle tasks like analyzing data loss prevention alerts and investigating insider threats. These agents can automate investigations and create tickets, but human analysts still control the final response. The tool aims to reduce alert noise and streamline workflows for teams dealing with too many alerts and stretched staffing.
WatchGuard introduces Rai AI digital worker for MSP security
WatchGuard launched Rai, an AI-powered digital worker that acts as an always-on security analyst for MSPs. Rai can monitor environments, correlate activity, and take pre-authorized actions in real time without human help. The tool is designed to handle first-response work, reducing after-hours alert handling and giving analysts more time for higher-value tasks. WatchGuard says Rai includes governance features so partners can control what actions it takes and see all its decisions.
Tips for parents setting healthy AI boundaries with teens
Artificial intelligence is now a common tool for teenagers, and parents need to set limits on its use. Lifestyle expert Meagan B. Murphy shared advice on where to draw those boundaries and how to put them into practice. The article emphasizes the importance of guiding teens on safe and responsible AI usage.
Lawsuit claims Perplexity shared chatbot chats with ad trackers
A new lawsuit called Doe v. Perplexity accuses the AI company Perplexity, along with Meta and Google, of sharing users' chatbot conversations for targeted advertising. The complaint says users were not warned that their messages were being sold via third-party tracking technologies, even in Incognito Mode. The case raises questions about how AI companies can make money while respecting user privacy. Legal experts say the claims could be avoided if companies were transparent and got informed consent from users.
AI labels meant to spot fakes are failing in practice
AI labels were supposed to help users identify fake content online, but they are not working well. The article highlights the case of Emily Hart, a fake influencer created by a student using Google Gemini, who gained thousands of followers. Platforms have rules requiring AI content to be labeled, but enforcement is difficult. The story shows how easy it is to create deceptive online personas using AI.
Meta sued by publishers and author Scott Turow over AI training
Five major publishers and bestselling novelist Scott Turow have sued Meta, alleging the company used copyrighted works to train its AI without permission. The complaint claims Meta's AI can mimic authors' styles and produce competing material. The lawsuit argues that AI-generated books are flooding marketplaces like Amazon and could push out human authors. The case highlights the ongoing debate between innovation and ownership in the age of AI.
Cursor's Composer autoinstall lets AI set up its own training environments
Cursor introduced Composer autoinstall, a system that uses AI to automatically set up complex development environments for training other AI models. The system works in two stages: first, a goal setting agent proposes commands and expected outputs, then a separate agent builds the environment. This approach is important for reinforcement learning training, which needs stable environments to work. Composer 2 scored 61.7 percent on a benchmark test, up from 47.9 percent for the previous version.
Marc Andreessen mocked for misunderstanding how AI chatbots work
Venture capitalist Marc Andreessen was ridiculed after a custom prompt he wrote for an AI chatbot was made public. In the prompt, he told the chatbot to never hallucinate or make anything up, showing a misunderstanding of how large language models work. Experts pointed out that hallucinations are a core issue with the technology and cannot be fixed by simply asking the AI not to do it. The incident raised questions about how well influential tech figures understand the AI they promote.
Proshort uses AI to coach sales teams in real time
Proshort is an AI-powered sales coach that provides real-time guidance by analyzing sales calls and conversations. The platform identifies where salespeople fall short and offers specific feedback to help them improve. Founder Gaurav Mishra says the tool has helped improve win rates by up to 20 percent and saves sales teams up to 10 hours per week on administrative tasks. Proshort has over 45 customers and crossed $1 million in annual recurring revenue within a year.
Federal and state rules for AI in health insurance claims reviewed
A new report examines how AI is being used in health insurance prior authorization and claims review. Insurers, providers, and patients all use AI in different ways during the claims process. The report highlights risks like inaccurate or biased outcomes and privacy breaches from AI tools. It also looks at the Trump administration's AI framework, which promotes AI development and recommends limited federal restrictions.
Sources
- AI Agents Now Take On SOC Grunt Work – Netskope Brings AgentSkope Into Security and Network Operations
- MSPs Get AI Workforce to Scale Managed Security
- Setting healthy boundaries with teens and artificial intelligence
- As Companies Monetize AI, Courts Will Weigh In
- AI labels were supposed to help users spot fakes. Here's why they're failing
- Meta sued for allegedly using copyrighted work to train AI
- Composer Autoinstall: AI Learns to Set Up Itself
- Marc Andreessen Mocked for Accidentally Revealing That He Seems to Have a Deep Misunderstanding of How AI Actually Works
- How Proshort is turning AI into a real-time sales coach
- Regulation of AI in Prior Authorization and Claims Review: A Look at Federal and State Consumer Protections
Comments
Please log in to post a comment.