This past year I hired, fired and then rehired an artificial intelligence scheduler called Amy. This is my first experience managing a machine. Sure, there was a novelty factor that was fun but she's also had a few West World episodes that were less fun. With so much media focus on how artificial intelligence and machines are replacing human knowledge workers, there is both excitementand fearabout the future of work. As intelligent technology can now interpret text, understand speech and make sense of images it has become possible to automate tasks that were previously the preserve of humans.

Today, most large-scale conversational AI agents (e.g. Alexa, Siri, or Google Assistant) are built using manually annotated data to train the different components of the system. Typically, the accuracy of the ML models in these components are improved by manually transcribing and annotating data. As the scope of these systems increase to cover more scenarios and domains, manual annotation to improve the accuracy of these components becomes prohibitively costly and time consuming. In this paper, we propose a system that leverages user-system interaction feedback signals to automate learning without any manual annotation. Users here tend to modify a previous query in hopes of fixing an error in the previous turn to get the right results. These reformulations, which are often preceded by defective experiences caused by errors in ASR, NLU, ER or the application. In some cases, users may not properly formulate their requests (e.g. providing partial title of a song), but gleaning across a wider pool of users and sessions reveals the underlying recurrent patterns. Our proposed self-learning system automatically detects the errors, generate reformulations and deploys fixes to the runtime system to correct different types of errors occurring in different components of the system. In particular, we propose leveraging an absorbing Markov Chain model as a collaborative filtering mechanism in a novel attempt to mine these patterns. We show that our approach is highly scalable, and able to learn reformulations that reduce Alexa-user errors by pooling anonymized data across millions of customers. The proposed self-learning system achieves a win/loss ratio of 11.8 and effectively reduces the defect rate by more than 30% on utterance level reformulations in our production A/B tests. To the best of our knowledge, this is the first self-learning large-scale conversational AI system in production.

A typical after-work scene at my house goes something like this. She chimes, then lights up. My husband says the persistent disconnect between me and Alexa is my fault--I need to pause more, speak more clearly, and maybe throw in a "please" now and then. But not long after she moved in--a necessary sidekick, I was told, to the new sound system he had installed--I started getting the feeling she preferred Bob over me, no matter how polite I was (although often I wasn't). Once she started piping up every time someone in the house called my name ("Alyssa!"),

Once upon a time, projectors were big, clunky, and, well, not exactly easy to move around if you want to watch the big game in your backyard or a movie at a friend's house. However, portable projectors have come a long way and Anker's Nebula Capsule II Smart Mini Projector is a solid bet. The Anker projector, which runs Android TV, is lightweight, works with Google Assistant, and has an auto-focus function that automatically adjusts the screen size to your projector screen within a second, not to mention other features that deliver a sharp and clear picture. I've been in the market for a smart projector so I can watch my favorite teams play and watch movies outside. I had high hopes of testing the Nebula Capsule II against a large projection screen in my backyard, but due to the high Florida humidity--and unbearable October heat--I decided to stick to testing the smart projector indoors. There are a few bare white walls in my house, so I set this projector up in my living room and got to testing.

Fox Business Briefs: Amazon is rolling out new tools to give users control over the stored voice recordings from their Alexa devices, amid a range of different privacy-related concerns. Voice assistants such as Amazon's Alexa, Apple's Siri and Google Assistant can be hacked by shining a laser on the devices' microphones, according to an international team of researchers. Dubbed "Light Commands," the hack "allows attackers to remotely inject inaudible and invisible commands into voice assistants," according to a statement from experts at the University of Electro-Communications in Tokyo and the University of Michigan. By targeting the MEMS (Microelectro-Mechanical Systems) microphones with lasers, the researchers say they were able to make the microphones respond to light as if it was sound. "Exploiting this effect, we can inject sound into microphones by simply modulating the amplitude of a laser light," they wrote in the research paper.

Gartner says artificial intelligence is expected to be common in the office by 2025, already seeing'huge pent-up demand' Artificial intelligence (AI) will be widely adopted in office environments in a variety of ways over the next few years as businesses invest in digital workplace initiatives, Gartner analysts have said. The trend is expected to gather steam as voice-activated personal assistants that have proved a hit at home begin to make inroads in the office. By 2025, the technology will "certainly be mainstream," said Matthew Cain, vice president and distinguished analyst at Gartner – even though privacy and security concerns have limited deployments so far. Cain was among the analysts who spoke at Gartner's Digital Workplace Summit in London. Gartner has separately predicted that consumer and business spending on smart speakers will pass $3.5 billion (€3.15 billion) in 2021, with 25% of digital workers using an AI assistant on a daily basis within the next two years.

When pressed on this point, Prasad emphasized that his team has made it easier for users to periodically auto-delete their data and opt out of human review. Neither option actually keeps the data from being used to train Alexa's myriad machine-learning models, though. In fact, Prasad alluded to ongoing research that would switch Alexa's training process to one where models can quickly be updated anytime there is new user data, more or less guaranteeing that the value from said data will be captured before it's disposed of. In other words, auto-deleting your data will mean only that it won't still be around to train future models once training algorithms have been updated; for current models, your data would be used in roughly the same way.

At its Ignite conference this year, Microsoft is continuing its A.I. push. By further integrating artificial intelligence into its Microsoft 365 applications -- including its Office suite products such as Outlook and Excel -- the company wants to offload some of the heavy human lifting to its Cortana digital smart assistant to help you stay productive. By bringing its A.I. smarts to Outlook on iOS, Microsoft is showing how you can truly go hands-free and eyes-free and still stay on top of your email. With the launch of a new Play My Emails feature, Outlook users on iPhone and iPad will be able to have Cortana read out emails using natural language technology. For commuters, this feature can save you time during your morning and evening commute, as Cortana will be able to read your email messages to you while you're driving, a move that places Microsoft's digital assistant in direct competition with music and audiobooks for your ear time.

Indeed, the legacy of in entertainment has conditioned us to think of it as technology that operates without human input. No wonder so many have been shocked to discover that Google Assistant relies on human help to improve its understanding of voice conversations or that numerous tech startups hire human workers to prototype and imitate functionality. The reality is that we are still far from achieving generalized that is functionally equivalent to the human mind. As the cofounder and CEO of a customer support automation platform that helps enterprises launch and train virtual agents, I've realized that whether the first generalized is born a year or 100 years from now, will always require human input and expertise -- technical and otherwise -- to operate at its full potential in a way that's ethical, responsible and safe. Below, we'll take a look at some examples of how relies on human input across a variety of established and emerging applications and explain why even the smartest will still require human assistance.

Alexa, Google Assistant, Siri, and Cortana can answer all sorts of questions that pop into users' heads, and they're improving every day. But what happens when a company like Amazon decides to crowdsource answers to fill gaps in its platform's knowledge? The result can range from amusing and perplexing to concerning. Alexa Answers allows any Amazon customer to submit responses to unanswered questions. When the web service launched in general availability a few weeks ago, Amazon gave assurances that submissions would be policed through a combination of automatic and manual review.