Close Menu
    Facebook X (Twitter) Instagram
    Facebook X (Twitter) Instagram
    The Gulf WeeklyThe Gulf Weekly
    • Home
    • UAE
    • KSA
    • GCC
    • Technology
    • Lifestyle
    • Sports
    The Gulf WeeklyThe Gulf Weekly
    Home»Technology»Microsoft, Google and xAI look to address AI regulation

    Microsoft, Google and xAI look to address AI regulation

    Editorial TeamBy Editorial TeamMay 12, 2026
    Facebook Twitter Pinterest LinkedIn Tumblr Reddit Telegram Email
    Share
    Facebook Twitter LinkedIn Pinterest Email

    Microsoft recently signed testing and evaluation agreements with the governments of the UK and U.S., the latter of which also sealed deals with Google DeepMind and xAI. The company went on to claim that AI regulation would be a boon for development of frontier systems. 

    The U.S. software giant used a blog to explain government scrutiny drives it to improve its advanced AI systems.

    “Well-constructed tests help us understand whether our systems are working as intended” and keeps the company alert to “risks, such as AI-driven cyberattacks and other criminal misuses”.

    Microsoft announced it is working with the AI Security Institute (AISI) in the UK and U.S. Centre for AI Standards and Innovation (CAISI) to assess its cutting-edge models and prepare protections.

    CAISI is a unit of the U.S. Department of Commerce which acts as the primary point of contact with the government for matters including testing, research and developing best-practices.

    It signed safety and evaluation deals with Anthropic and OpenAI in 2024.

    Spat

    The U.S., in particular, sparked headlines in recent weeks in a well-publicised spat with Anthropic, so the latest moves by each government appear timely.

    CAISI’s arrangements grant it early access to Microsoft, Google DeepMind and xAI’s models which director Chris Fall said is “essential” to perceiving the potential dangers posed by their technologies. He said the deals come at a “critical moment”.

    The companies committed to work on improving the US government’s understanding of what AI can do, along with potential national security risks and the level of global competition.

    CAISI clocked a tally of at least 40 evaluations to-date.

    In the 2025 edition of its Frontier AI Trends Report, AISI noted “longer, more sophisticated attacks” were required to jailbreak AI models with the most rigorous protections “for certain malicious request categories”, though cautioned the “efficacy of safeguards varies between models” and no system tested was free of vulnerabilities.

    Source: Mobile World Live

    Image Credit: Stock Image


    Source: Tahawul Tech

    Related Posts

    Sharjah Digital Department, Shams AI Club train staff on AI for media, marketing

    July 30, 2026

    Lenovo powers EWC as founding partner with latest Legion desktop PCs, monitors

    July 30, 2026

    RTA conducts comprehensive campaign to verify safety of Dubai Metro tunnel ventilation systems

    July 29, 2026
    Don't Miss

    Salman Rushdie attacker convicted of terror offenses

    KSA July 30, 2026

    NEW YORK — The US man who stabbed and partially ​blinded author Salman Rushdie at…

    Inside Dubai’s humanitarian hub helping stranded migrants find their way home

    July 30, 2026

    flydubai strengthens operations in Italy with increased frequency to Naples, Milan

    July 30, 2026

    Three Greek firefighters die as wildfires threaten parts of southern Europe

    July 30, 2026
    Our Picks

    Salman Rushdie attacker convicted of terror offenses

    July 30, 2026

    Inside Dubai’s humanitarian hub helping stranded migrants find their way home

    July 30, 2026

    flydubai strengthens operations in Italy with increased frequency to Naples, Milan

    July 30, 2026

    Three Greek firefighters die as wildfires threaten parts of southern Europe

    July 30, 2026
    2026. All rights reserved.
    • UAE
    • KSA
    • GCC
    • Technology
    • Lifestyle
    • Sports
    • Contact us

    Type above and press Enter to search. Press Esc to cancel.