Microsoft has said it is "deeply sorry" for the racist and sexist Twitter messages generated by the so-called chatbot it launched this week, after the artificial intelligence program went on an embarrassing tirade. The bot, known as Tay, was targeted at 18 to 24-year-olds in the US and was designed to become "smarter" as more users interacted with it. Instead, it quickly learned to parrot a slew of anti-Semitic and other hateful invective that human Twitter users started feeding the program, forcing Microsoft Corp to shut it down. Following the setback, Microsoft said in a blog post it would revive Tay only if its engineers could find a way to prevent web users from influencing the chatbot in ways that undermine the company's principles and values. "We are deeply sorry for the unintended offensive and hurtful tweets from Tay, which do not represent who we are or what we stand for, nor how we designed Tay," wrote Peter Lee, Microsoft's vice president of research.

Marketers use the word "handmade" to signal that a product is worth an extra couple bucks -- be it cheddar or loafer. But it's not just physical goods that are made more quickly and cheaply by machine. Artificial intelligence is making decisions that would previously be done by humans. In a new post, Fast Company's Sarah Kessler reports on the on-demand concierge startup GoButler, which just went from having humans handle requests to solely relying on algorithms. "My general view is that people will always want convenience, but they're not willing to pay premiums for it," says GoButler CEO Navid Hadzaad.

DARPA has a new Grand Challenge underway, but it's not an automationBig Performance from Low-Power Hardware. Read more ... » moonshot like the selfWhen is CES 2016, what will be the best gadgets and how much are ticket prices?. Read more ... »-driving car challenges of the earlyHome IoT security could come from a glowing rock next year. The Defense Department's R&D wing wants to revolutionize something with a bit less sex appeal, but plenty of real-world applicationsNvidia Releases Machine Learning Products for Hyperscale Datacenters. The Spectrum Collaboration Challenge, which DARPA has cleverly abbreviated SC2, is about getting the billions and billions of wireless devices out there to play nice together rather than fight for space in the increasingly crowded RF landscape.

A user told Tay to tweet Trump propaganda; she did (though the tweet has now been deleted). Microsoft has apologized for the conduct of its racist, abusive machine learning chatbot, Tay. The bot, which was supposed to mimic conversation with a 19-year-old woman over Twitter, Kik, and GroupMe, was turned off less than 24 hours after going online because she started promoting Nazi ideology and harassing other Twitter users. The company appears to have been caught off-guard by her behavior. A similar bot, named XiaoIce, has been in operation in China since late 2014.

DARPA has a new Grand Challenge underway, but it's not an automation moonshot like the self-driving car challenges of the early 2000s or the recent (and hilarious) Robotics Challenge. The Defense Department's R&D wing wants to revolutionize something with a bit less sex appeal, but plenty of real-world applications: radio frequency spectrum splitting. The Spectrum Collaboration Challenge, which DARPA has cleverly abbreviated SC2, is about getting the billions and billions of wireless devices out there to play nice together rather than fight for space in the increasingly crowded RF landscape. Seriously, check out that cool chart at the top (much bigger version at DARPA's site if you want to print out the poster – 56k warning): everything is spoken for right up to the border with microwave frequencies, and more gadgets are crowding into each one daily. "The current practice of assigning fixed frequencies for various uses irrespective of actual, moment-to-moment demand is simply too inefficient to keep up with actual demand and threatens to undermine wireless reliability," said William Chappell, director of DARPA's Microsystems Technology Office, in a DARPA press release.

Artificial Intelligence, or AI, is expected to be the next big thing in technology, and many tech companies are investing heavily to make sure they don't lose out on opportunities as and when the technology becomes more mainstream in the coming years. While AI has already started intriguing and exciting industry insiders and tech enthusiasts alike, the increasing chatter surrounding the new technology has given rise to multiple concerns regarding job losses and science fiction-like scenarios of AI-enabled robots starting to exert control over humans, eventually taking over the planet, not unlike the dystopian sci-fi movie'I, Robot'. While Alphabet Chairman, Mr. Eric Schmidt, has already admitted that jobs may be at stake because of artificial intelligence, the founder and CEO of Facebook, Mr. Mark Zuckerberg, recently categorically denied that AI-enabled robots are about to take over the planet any time soon. Whatever be the case, most established technology companies are diving into AI-related research headlong, and while American tech giants like Google and Facebook are already burning the midnight oil on research related to artificial intelligence, South Korean conglomerate, Samsung, is also spending the big bucks by investing in a number of tech startups worldwide, in an effort to increase its presence in the sector. According to reports in the South Korean media, Samsung Venture Investment, which happens to be a subsidiary of the Samsung Group and hence, a sister concern for Samsung Electronics, has already invested a truckload of money in more than ten global tech companies, all of which are deeply involved in AI-related research.

Microsoft has apologized for the conduct of its racist, abusive machine learning chatbot, Tay. The bot, which was supposed to mimic conversation with a 19-year-old woman over Twitter, Kik, and GroupMe, was turned off less than 24 hours after going online because she started promoting Nazi ideology and harassing other Twitter users. The company appears to have been caught off-guard by her behavior. A similar bot, named XiaoIce, has been in operation in China since late 2014. XiaoIce has had more than 40 million conversations apparently without major incident.

Everyone in business wants to know what's going to happen in the future, and for some time now Ford has been investing in futurism, an evolving academic and professional discipline. The need for this was particularly evident after the Business Insider Transportation team in New York spent a few days at the New York Auto Show, asking everyone to predict was will happen in 2016 -- and beyond. The car business these days is all about change: automakers becoming "mobility providers," electric cars potentially displacing gas-powered vehicles, even autos driving themselves. Heck, even Apple may get in on the action. For nearly a decade, Sheryl Connelly has been Ford's manager of global consumer trends and futuring.

A fully loaded fireproof mini tankbot has proved size really doesn't matter. Called Uran-9, this mini tank bot stands just a few feet taller than a human and is fully loaded with a machine gun, missiles and a 30-millimeter cannon that fires 350 to 400 rounds per minute. A Russian defense organization created this unmanned vehicle to provide reconnaissance and fire support to infantry and counter-terror units. Uran-9 stands just a few feet taller than the average human being, but there is no need to be any bigger as the machine does not transport soldiers. This vehicle will assists infantry units and counter-terrorism groups by reaching places soldiers are unable to travel.

We investigate the choice of tuning parameters for a Bayesian multi-level group lasso model developed for the joint analysis of neuroimaging and genetic data. The regression model we consider relates multivariate phenotypes consisting of brain summary measures (volumetric and cortical thickness values) to single nucleotide polymorphism (SNPs) data and imposes penalization at two nested levels, the first corresponding to genes and the second corresponding to SNPs. Associated with each level in the penalty is a tuning parameter which corresponds to a hyperparameter in the hierarchical Bayesian formulation. Following previous work on Bayesian lassos we consider the estimation of tuning parameters through either hierarchical Bayes based on hyperpriors and Gibbs sampling or through empirical Bayes based on maximizing the marginal likelihood using a Monte Carlo EM algorithm. For the specific model under consideration we find that these approaches can lead to severe overshrinkage of the regression parameter estimates in the high-dimensional setting or when the genetic effects are weak. We demonstrate these problems through simulation examples and study an approximation to the marginal likelihood which sheds light on the cause of this problem. We then suggest an alternative approach based on the widely applicable information criterion (WAIC), an asymptotic approximation to leave-one-out cross-validation that can be computed conveniently within an MCMC framework.