Google Responds to Training Accusations

Google publicly states that it does not use Gmail messages, attachments, or private email content to train its Gemini AI models, calling circulating claims “misleading.”

Confusion Around Smart Features

The issue stems from Gmail’s Smart Features, like Smart Compose and automated reminders, which have existed for years but were mistaken by users as AI training permissions. Google says these features personalize the inbox but don’t send email content for model training.

Reports of Settings Being Reset

Some users, including a Verge staff member, reported that Smart Features they had previously turned off were suddenly switched back on. Google attributes this to a January update that separated personalization settings across Workspace and other products.

The Opt-Out Process Is Split in Two Places

Users must turn off Smart Features inside Gmail and then separately disable “Smart Features and Personalization” under Google Account settings. This two-step process is contributing to confusion about whether email data is being used.

Ongoing Privacy Concerns

Even with Google’s denial, many users remain uneasy about transparency and data handling practices, raising broader questions about how clearly companies communicate consent and personalization options.

Nostalgia by algorithm

Users of Sora are generating streams of dreamy 80s/90s-style content, sun-drenched cul-de-sacs, big hair, vintage cars, and “MTV era” teens, produced by AI models with little regard to historical accuracy or context. 

Deep-fake graveyard of public figures

The app has churned out bizarre clips of deceased figures like Fred Rogers rapping with rap stars, or Stephen Hawking starring in extreme stunts, highlighting how quickly AI can create hyper-realistic but false likenesses. 

Virality over meaning

While these videos get views, the article argues they reflect a low-effort “slop machine” mindset, content optimized for clicks rather than originality, nuance, or lasting value. 

Legacy and consent under threat

With public figures’ likenesses being used without meaningful control, legal and ethical questions mount around who owns our image and memory when AI can resurrect them for memes. 

The bigger picture: AI output and attention economics

This trend suggests that while AI has scaled content creation, it hasn't yet elevated cultural value. The moment-to-moment novelty may fade, especially if what’s produced lacks depth or substance. 

Define Your Event Goals & Requirements

Start by outlining the event type (conference, workshop, wedding, expo), expected audience size, key activities, and time constraints. Provide this information to an AI planner (like Notion AI, ClickUp AI, or ChatGPT) so it can generate a realistic structure.

Use AI to Draft an Initial Schedule

Feed your agenda details into an AI tool to automatically create a draft timeline. AI can allocate ideal time slots for sessions, breaks, transitions, speaker sets, and attendee flow. This gives you a ready-made skeleton schedule within seconds.

Optimize Timing & Flow with AI Suggestions

Ask the AI to evaluate bottlenecks, overlaps, or long gaps in your schedule. Many tools can adjust session durations, ensure smooth attendee movement, and create efficient sequencing (e.g., keynote → breakout → networking).

Automate Speaker & Resource Coordination

Upload speaker info, room availability, and equipment needs to your AI system. It can automatically match speakers to rooms, avoid conflicts, and send reminders, reducing manual coordination. Tools like Calendly AI or Motion can also auto-sync calendars.

Generate Shareable Schedules & Real-Time Updates

Use AI to export polished versions of your event schedule in formats like PDFs, mobile agendas, or interactive timelines. On event day, AI can update the schedule in real time if sessions run late or last-minute changes occur, ensuring smooth event management.

Orum

Orum is an AI-powered live conversation platform built for sales teams to optimize outbound calling. Its parallel-dialer can call multiple numbers simultaneously, intelligently skip voicemails or disconnected lines, and connect reps only when a prospect answers. On top of that, Orum offers real-time AI coaching, call-roleplay agents, and analytics to improve rep performance. By automating dialing and enhancing call quality, Orum helps sales teams generate far more live conversations, spend less time on manual tasks, and scale pipeline growth. 

Hoxhunt

Hoxhunt is a human risk management platform that uses AI to strengthen corporate security culture and reduce cyber risk. Through adaptive phishing simulations and security-awareness training, it personalizes the experience for each employee, making learning continuous and behavior-driven. The platform’s AI identifies risky behaviors, automatically triggers micro-training, and provides analytics to measure how effectively employee risk is changing over time. 

Absorb LMS

Absorb LMS is an AI-enhanced learning management system for businesses to deliver personalized training, compliance education, and upskilling. It's built-in generative AI (Absorb Create) helps course creators draft interactive lessons, quizzes, and assessments in minutes. For learners, the platform tailors learning paths based on skills, interests, and progress. It also automates admin tasks and provides smart recommendations to reduce manual work and maximize engagement. 

EvenUp (AI Demand Packages)

EvenUp provides AI-powered demand-letter drafting tools tailored for personal injury (PI) law firms. Their platform transforms raw case files, medical records, treatment data, and bills into polished demand packages, including medical chronologies, narrative summaries, and evidence-backed arguments. By combining AI with legal expertise, EvenUp’s solutions increase settlement value, speed up demand generation timelines, and reduce manual drafting overhead. 

Lair East Labs

Lair East Labs is a New York–based startup accelerator that empowers early-stage founders to scale internationally, particularly between the U.S. and Asian markets. Through its 3-month program, Lair East Labs provides funding (up to ~$150K per startup), mentorship, and operational support in legal, product, marketing, and hiring. The accelerator’s network helps founders navigate cross-border growth and build global businesses. 

Altan

Altan is an “agent-native” platform that uses AI to design, build, and run software autonomously by orchestrating specialized agent squads. Users simply describe their idea,  through text or voice, and Altan summons AI agents with roles like full-stack engineers, UX designers, and product managers to create production-ready software. The platform doesn’t just stop at development: Altan also deploys and operates the software, making it possible for apps to self-manage. Based in Barcelona, the startup recently raised $2.5 million in a pre-seed round to scale its agent-native vision.

Bem (bem.ai)

Bem is building an AI infrastructure layer that automates unstructured, enterprise workflows, such as data trapped in PDFs, emails, and spreadsheets. Rather than just extracting information, bem’s platform provides trainable AI “functions” that orchestrate entire end-to-end business processes. By doing so, it helps companies in highly regulated or data-intensive industries (like insurance, supply chain, and healthcare) streamline operations and reduce manual overhead.  

ForkOff

ForkOff is a Web3 + AI marketing & retention company that helps blockchain-native and AI-first brands drive engagement, recall, and sales. According to its Wellfound listing, ForkOff works with Web3 teams to build community, handle outreach, and manage retention, blending AI-driven marketing strategies with Web3 insights. Their team is small (1-10 employees), which suggests they are early but nimble, focused on high-growth partnerships in the crypto + AI space.