Language Selection

English French German Italian Portuguese Spanish

Google's do no evil AI style likely to clash with open source approach

Filed under
Google
OSS
Sci/Tech
  • Google's do no evil AI style likely to clash with open source approach

    Google outlined its artificial intelligence principles in a move to placate employees who were worried about their work and research winding up in U.S. weapons systems.

    Guess what? It's already too late. There's no way that Google's open source approach and its headline principle to not allow its AI into weapons is going to mesh. Chances are fairly good that the technology already open sourced is in some fledgling weapon system somewhere. After all, TensorFlow and a bunch of other neural network tools are pretty damn handy.

  • Read Google's AI ethics memo: 'We are not developing AI for use in weapons'
  • Google Plans Not to Renew Its Contract for Project Maven, a Controversial Pentagon Drone AI Imaging Program
  • Google promises not to use A.I. for weapons or surveillance, for the most part
  • Google pledges not to develop AI weapons, but says it will still work with the military

    Google has released a set of principles to guide its work in artificial intelligence, making good on a promise to do so last month following controversy over its involvement in a Department of Defense drone project. The document, titled “Artificial Intelligence at Google: our principles,” does not directly reference this work, but makes clear that the company will not develop AI for use in weaponry. It also outlines a number of broad guidelines for AI, touching issues like bias, privacy, and human oversight.

    While the new principles forbid the development of AI weaponry, they state that Google will continue to work with the military “in many other areas.” Speaking to The Verge, a Google representative said that had these principles been published earlier, the company would likely not have become involved in the Pentagon’s drone project, which used AI to analyze surveillance footage. Although this application was for “non-offensive purposes,” and therefore hypothetically permitted under these guidelines, the representative said it was too close for comfort — suggesting Google will play it safe with future military contracts.

More in Tux Machines

Fedora 30 Workstation review - Smarter, faster and buggier

Fedora 30 is definitely one of the more interesting releases of this family in a long-time. It brings significant changes, including solid improvements in the desktop performance and responsiveness. Over the years, Fedora went from no proprietary stuff whatsoever to slowly acknowledging the modern needs of computing, so now it gives you MP3 codecs and you can install graphics drivers and such. Reasonable looks, plus good functionality across the board. However, there were tons of issues, too. Printing to Samba, video screenshot bug, installer cropped-image slides, package management complications, mouse cursor lag, oopses, average battery life, and inadequate usability out of the box. You need to change the defaults to have a desktop that can be used in a quick, efficient way without remembering a dozen nerdy keyboard shortcuts. All in all, I like the freshness. In general, it would seem the Linux desktop is seeing a cautious revival, and Fedora's definitely a happy player. But there are too many rough edges. Well, we got performance tweaks after so many years, and codecs, we might get window buttons and desktop icons one day back, too. Something like 6/10, and definitely worth exploring. I am happy enough to do two more tests. I will run an in-vivo upgrade on the F29 instance on this same box, and then also test the distro on an old Nvidia-powered laptop, which will showcase both the support for proprietary graphics (didn't work the last time) and performance improvements, if they scale for old hardware, too. That's all for now. Read more

Events: Automotive at LF, Linux Clusters Institute, Linux Plumbers Conference (LPC)

  • Automotive Linux Summit and Open Source Summit Japan Keynote Speakers and Schedule Announced
    The Linux Foundation, the nonprofit organization enabling mass innovation through open source has announced the speaker line up for Open Source Summit Japan and Automotive Linux Summit. One registration provides access to all content at both events, which will be held July 17-19 at the Toranomon Hills Forum in Tokyo. Open Source Summit Japan (OSSJ) and Automotive Linux Summit (ALS) will bring together top talent from companies on the leading edge of innovation including Toyota Motor Corporation, Uber, Intel, Sony, Google, Microsoft and more. Talks will cover a range of topics, with ALS talks on everything from infrastructure and hardware to compliance and security; and OSSJ sessions on AI, Linux systems, cloud infrastructure, cloud native applications, open networking, edge computing, safety and security and open source best practices.
  • Register Now for the 2019 Introductory Linux Clusters Institute Workshop
    Registration is now open for the 2019 Linux Clusters Institute (LCI) Introductory Workshop,which will be held August 19-23, 2019 at the Rutgers University Inn & Conference Center in New Brunswick, NJ. This workshop will cover the fundamentals of setting up and administering a high-performance computing (HPC) cluster and will be led by leading HPC experts.
  • Additional early bird slots available for LPC 2019
    The Linux Plumbers Conference (LPC) registration web site has been showing “sold out” recently because the cap on early bird registrations was reached. We are happy to report that we have reviewed the registration numbers for this year’s conference and were able to open more early bird registration slots. Beyond that, regular registration will open July 1st. Please note that speakers and microconference runners get free passes to LPC, as do some microconference presenters, so that may be another way to attend the conference. Time is running out for new refereed-track and microconference proposals, so visit the CFP page soon. Topics for accepted microconferences are welcome as well.

Security Leftovers

  • Security updates for Wednesday
  • Illumos-Powered OmniOS Gets Updated Against MDS / ZombieLoad Vulnerabilities
    While it was just earlier this month that the OpenSolaris/Illumos-based OmniOS saw a big LTS release, it's already been succeeded by a new release given the recent Intel MDS / Zombieload CPU vulnerabilities coming to light. There are new spins of OmniOS for all supported releases. These new OmniOS Community Edition releases mitigate against the Multiarchitectural Data Sampling (MDS) vulnerabilities and also bundle in the updated Intel CPU microcode.
  • Hackers Hack A Forum For Hacked Accounts: Here’s How
    A group of hackers failed to deploy security mechanisms to secure the storage where they store hacked accounts and another hacker group hacked it. The story is indeed funny and real. Infamous forum named OGUSERS which is popular amongst hackers for obtaining “OG” Instagram, Twitter usernames, hacked accounts of Domino’s Pizza, Steam, PlayStation Network, and other online accounts was hacked by a hacker group and its data was published in another hacker forum.
  • Security Announcement: Disabling SMT by default on affected Intel processors
    This is an important announcement with an upcoming change in the next Core Update of IPFire. Because of the recent vulnerabilities in Intel processors, the IPFire team has decided, that - to keep systems as secure as possible - Simultaneous Multi-Processing (SMT) is automatically disabled if the processor is vulnerable to one of the attacks. SMT is also called Intel(R) Hyper-Threading Technology and simulates more virtual cores than the system has. This allows to perform faster processing when applications benefit from it. Unfortunately with networking, we benefit from that. Therefore the effect of disabling SMT will be a very signifiant performance impact of around 30% or more. Applications that will be affected in IPFire are the firewall throughput itself as well as other CPU and memory-bound tasks like the web proxy and the Intrusion Prevention System. On systems that are not vulnerable for this attack, SMT is being left enabled. If you still want to disable it, please do so in the BIOS of your firewall.

Android Leftovers