Skip to content
IFAI

AI safety & security

Industry tag1,497 stories

  1. IndustryReach 49

    US Patent Office says AI can help, but only humans can be inventors

    The USPTO just issued new guidance declaring that only "natural humans" can be listed as inventors on patents, ruling out AI models as inventors or co-inventors. While AI-assisted inventions aren't barred from patent protection, human applicants must prove they made a "significant contribution" to the conception—such as specific prompt construction or heavily modifying the outputs. Simply running a prompt or owning the model doesn't cut it.

  2. IndustryReach 36

    Men are using AI to write love letters—and people hate it

    McAfee just revealed that 45% of men plan to use AI to draft Valentine’s Day messages this year, up from 30% last year. While 69% say AI-crafted charm gets better responses, 57% admit they'd be hurt or offended if they found out code wrote their love note.

  3. IndustryReach 48

    Protesters picket OpenAI office over military ties and AGI fears

    Dozens of activists from Pause AI and No AGI just rallied outside OpenAI's San Francisco headquarters, demanding the lab halt work on AGI and urging engineers to quit. The protest followed OpenAI's quiet removal of a ban on military use from its policies and its subsequent work with the Pentagon. Organizers argue tech labs cannot self-regulate and need strict external oversight.

  4. IndustryReach 35

    OpenAI is building an AI agent that takes over your screen

    OpenAI is reportedly working on an autonomous AI agent capable of taking control of user devices to handle tasks directly, according to The Information. Instead of just answering prompts, it could click, type, and transfer data across apps for you. Why it matters: bringing JARVIS-style automation to everyday devices sounds incredible, but handing full device access to cloud AI creates serious privacy and security hurdles.

  5. IndustryReach 66

    The FCC outlaws AI voices in robocalls

    The FCC just banned the use of AI-generated voices in robocalls, ruling that voice-cloning tech falls under existing federal consumer protection laws. State attorneys general now have the authority to prosecute robocallers solely for deploying AI voices, giving regulators immediate legal teeth after high-profile political deepfakes.

  6. IndustryReach 59

    OpenAI builds a child safety team to protect young users

    OpenAI just launched a dedicated child safety team and is hiring enforcement specialists to prevent minors from misusing its AI. The group will work alongside legal and policy teams to review sensitive underage content and handle safety incidents.

  7. AI toolsReach 39

    How to keep your art out of AI generators

    Artists now have a few ways to push back against generative AI scraping their work without consent. While tech giants like OpenAI and Meta offer cumbersome opt-out forms that don't touch existing models, tools like Glaze and Nightshade let creators cloaking or actively poison their images. Why it matters: Until copyright laws catch up, pixel-level sabotage is one of the few proactive defenses visual artists have left.

  8. AI toolsReach 52

    OpenAI adds deep watermarks across text, audio, and images

    OpenAI just rolled out a multi-layered provenance system across its AI outputs, pairing C2PA metadata with invisible SynthID and textGrain watermarks. It also launched a public verification portal so anyone can check whether an asset came from its models. Why it matters: embedding signals directly into pixels and token choices makes AI origins much harder to scrub off.

  9. IndustryReach 34

    YouTube's 2024 vision puts AI front and center

    YouTube CEO Neal Mohan just outlined the platform's four big bets for 2024, placing generative AI right at the top. Mohan emphasized that AI should empower creators rather than replace them, spotlighting tools like Dream Screen for Shorts. Crucially, as deepfakes loom over global elections, YouTube will roll out transparent labels for realistic synthetic content in the coming months.

  10. IndustryReach 42

    Meta starts labeling AI-generated images across all its apps

    Meta just announced it will start labeling AI-generated images on Facebook, Instagram, and Threads in the coming months. The system will detect invisible markers based on C2PA and IPTC standards from tools made by OpenAI, Google, Adobe, and others. For AI video and audio where cross-industry signals are lacking, Meta will require users to self-disclose under threat of penalties.

  11. IndustryReach 38

    The UK's $125M AI bet skips new laws

    The UK government just committed over £100 million ($125M) to bolster homegrown AI research and upskill regulators, while officially confirming it won't rush into new AI legislation. Instead of copying the EU’s strict statutory framework, Britain is betting that flexible, sector-specific oversight will keep AI builders happy on home turf.

  12. IndustryReach 36

    Microsoft teams up with Semafor to reinvent news research

    Microsoft and OpenAI just partnered with news startup Semafor to launch "Signals," a multi-source breaking news feed. The twist? AI is strictly used as a research assistant, while human journalists still do all the actual reporting and writing. Why it matters: After getting sued by The New York Times, Microsoft is trying to show it can work with newsrooms instead of replacing them.

  13. IndustryReach 44

    Scammers use deepfake CFO in a video call to steal $25M

    A multinational firm just lost $25.6 million (HK$200 million) in Hong Kong after an employee was duped by a deepfake group video call. Scammers recreated the UK-based CFO and other colleagues using public footage to order 15 secret bank transfers. Why it matters: It marks an alarming escalation in enterprise fraud—you can no longer trust your eyes on a company call.

  15. IndustryReach 42

    FCC moves to make AI voice cloning in robocalls illegal

    The FCC just proposed officially declaring AI voice cloning in robocalls illegal under existing consumer protection laws. Spurred by deepfaked calls mimicking President Biden in New Hampshire, the ruling would treat cloned voices as prohibited artificial speech, making it much easier for regulators to prosecute fraudsters without waiting to prove secondary crimes.

  16. Open sourceReach 56

    Mistral's accidental GPT-4 rival

    Mistral CEO Arthur Mensch just confirmed that the mystery open-source model “miqu-1-70b”—which scored near GPT-4 on benchmarks after surfacing on 4chan and Hugging Face—is indeed an old Mistral checkpoint retrained from Llama 2, leaked by an early customer’s employee. Why it matters: it proves open-source AI is closing the frontier performance gap much faster than expected.

  17. IndustryReach 37

    The New York Times builds an in-house AI skunkworks

    The New York Times just started assembling a dedicated newsroom team to explore prototyping generative AI tools for reporting and storytelling. Led by editorial director Zach Seward, the hybrid crew of engineers, designers, and editors will test internal workflows. Why it matters: Even while locked in a lawsuit with OpenAI, the paper knows it can't ignore the tech—though it insists human journalists will still write the news.

  18. IndustryReach 41

    Crypto meets AI: Where it works and where it breaks

    A technical blog post just dissected the real-world intersections of crypto and AI, pointing out where the hype meets practical utility. The most promising applications involve AIs acting as players in micro-prediction markets or smart contract assistants, but using models to enforce governance rules exposes them to severe adversarial attacks. Why it matters: Decentralization and open models clash with machine learning vulnerabilities, meaning not all crypto-AI mashups are built to last.

  19. IndustryReach 43

    Fake Biden robocall tells voters to stay home

    A fake robocall mimicking President Biden just urged New Hampshire voters to skip the primary, and experts say it is almost certainly an AI deepfake. Media detection firm Reality Defender confirmed the audio was likely manipulated. Anyone can now clone political figures with minimal audio and cheap tools, highlighting a massive gap in election regulations.

  20. AI toolsReach 63

    Nightshade lets artists poison AI scrapers

    Researchers at the University of Chicago just released Nightshade v1.0, a free desktop tool designed to fight unauthorized AI training. It subtly alters pixels so that while images look normal to human eyes, an AI scraping them sees something totally different—like a cow morphing into a handbag. The goal is to make scraping uncorrupted data too costly to ignore.