Category / AI / visualization

    Loading posts...
  • Instagram now flags manipulated images

    As reported by PetaPixel, Instagram is now using a combination of AI and humans to flag manipulated images. Photoshop has been a verb for quite some time, but the move into the “post-evidence” world, the issue of provenance is rapidly becoming a challenge. Since PetaPixel is a photography-focused site, they highlight that there is a whole…

  • 21st Century Manhattan Project

    That was the term I used for virtual humans back in 2013 during a TedX talk “Analog Soul – Digital World.” I argued that avatars would progress to the point where their impact would reach far beyond just their intended use. Looks like we’re one step closer to that, courtesy of Neon. Part of the…

  • will trade secrets kill the 6th amendment?

    As we move into an increasingly AI-influenced world, we will have to continually revisit our governance structures and approaches to figure out what freedoms still exist, what are at risk, and what need to be rethought. The recent move to include algorithms in the bail bond process should lead us to consider potential impacts to…

  • triple Uber failure

    photo by Uber We knew this day would come. The Tesla autopilot incident a few years back was a harbinger, but since a human was still “in control” it really wasn’t the test that autonomous vehicles had coming. But the AZ incident is the first case of a pedestrian being killed by a robocar. The…

  • running blades

    Blade Runner 2049, while not crushing the box office, certainly has some people excited (my artist/wife being one of them). We saw it last night and my 3 word review is: Brilliant, flawed, beautiful It did provide a new tag line for me to use – “if it’s real, it should be messy.” The context…

  • no screens?

    I don’t necessarily agree with all of the points, but the new immersive mediums (NIMs) will profoundly what we think of as a “screen”, and AI will certainly replace some visual/tactile interfaces.

  • ai for good

    MIT has a new AI system that will detect 85% of cyber attacks. Actually, that is pretty good performance, and fairly useful as often you don’t know you’re under attack until it is too late. Score one for the bots (that aren’t selling you something).

  • corrupted by twitter

    Or at least it seems that way. Microsoft’s AI twitter-bot went from nice and innocent to raging racist in less than 24 hours. Now where did I leave my self-driving car? I hope it isn’t watching Mad Max films…

  • doc in a box not

    Skip Rizzo’s group at ICT has been working with VR for treating PTS for some time now. He always emphasizes that our systems are not a “doc in a box” – it requires a trained clinician in the loop. Of course digital healthcare is coming, and in fact that’s what we talked about at SxSW…