Advances in artificial intelligence 2018

artificial intelligence innovations 2018 and innovation management artificial intelligence
EmmaGoulding Profile Pic
EmmaGoulding,Vatican City,Professional
Published Date:06-07-2017
Your Website URL(Optional)
Comment
Artificial Intelligence Innovation Report Powered by 2016AI Artificial Intelligence Innovation Report 2016 Introduction Once the preserve of science AlphaGo eclipsed the 1997 Natural Language processing fiction, Artificial Intelligence achievement’s of IBM’s Deep is employed to facilitate human (AI) is now firmly embedded in Blue when it defeated Lee Sedol interaction. our day-to-day lives. Maps that in a five-game match of Go. Siri compute the optimal driving and Cortana are AI assistants we This AI is narrow because it route for a journey based on can carry in our pockets. And yet is incapable of carrying out current traffic, smart song these achievements only hint at any task outside of its specific suggestions from Pandora, the future of AI. purpose. Impressive as AlphaGo friend recommendations, job may be at playing Go, it would recommendations, product To understand the current be flummoxed if you asked it for recommendations based on state of AI and what the future restaurant recommendations. your purchase history — all holds, we must first split our are powered by AI. Even email understanding of AI into three Artificial General Intelligence spam filters, which learn to definitions: (AGI): Sometimes referred to as differentiate between email you Strong AI or Human-Level AI. At do and don’t want to receive, Artificial Narrow Intelligence this level, AI would be capable rely on a form of AI. (ANI): Sometimes referred to as of performing any task to the Weak AI, this is the most limited standard of a human. Precisely In recent years we’ve become form of AI and this is what is when AGI will be achieved is a accustomed to seeing AI currently possible in 2016. subject of debate, with most AI referenced in business tag ANI is capable of carrying out specialists predicting it will occur lines and popular culture. In specific tasks brilliantly, using around 2040. 2011 IBM Watson defeated the a combination of advanced world’s two best players of the algorithms, deep learning To reach AGI requires hardware Jeopardy quiz show. In March and various other techniques comparable to the human 2016 Google DeepMind’s depending on the use. Often, brain, which can process 2 Powered byAI Artificial Intelligence Innovation Report 2016 Introduction approximately 10 quadrillion to develop improvements on its end to global warming, a cure calculations per second on own AI. This would enable it to to cancer, human extinction, just 20 watts of power. Such become progressively smarter, the end of human mortality, and a computer actually already and to constantly build upon eternal agelessness are all being exists in China’s Tianhe-2, its own intelligence with an considered as very possible which can handle 34 quadrillion exponential increase in gains. outcomes. Stephen Hawking calculations per second. Spread describes the advent of ASI as across 720 square meters, Artificial Superintelligence a moment that “could spell the costing USD 390 million, and (ASI): This is an AI that is end of the human race”, while requiring 24 megawatts of more advanced than human Ray Kurzweil believes it could power, Tianhe-2 is clearly intelligence. Due to the Law of introduce a human-AI synthesis inappropriate for the consumer Accelerating Returns, this state or technological singularity that market. However, it does could potentially be realized in can best be described as a form demonstrate that hardware is mere days of even hours after of utopia. unlikely to be the biggest block AGI is achieved. on the development of AGI. This may sound like science The consequences of fiction, but if we look at the ANI Developing software that is as developing ASI are, by its very of 2016 we can already see the advanced as the human mind definition, unknown. This level impact of these developments, poses a greater challenge. of intelligence would soon most noticeably on the While such software does not be many billion times more workforce. currently exist, the majority of intelligent than the smartest computer scientists believe the human. Because of this, it is It is estimated that 50% of most promising approach could impossible for anyone to predict today’s jobs across the globe be to build an AI that can code with certainty what this leap in will be serviced by AI in 2050. changes into itself — enabling it technology will bring forth. An Many of the ANI examples 3 Powered byAI Artificial Intelligence Innovation Report 2016 Introduction below already demonstrate this a picture of a dog and a cat— adapted the natural environment movement, with repetitive, data- that remain incredibly complex for them. X.ai may be able to driven, non-social occupations for our current generation of take the pain out of scheduling being the most susceptible to ANI. As computer scientist meetings, but it can only do so early replacement. Roles that Donald Knuth explains, “AI has if everyone keeps their digital require lateral, creative thinking, by now succeeded in doing calendar up to date and attaches empathy or social sensitivity are essentially everything that the necessary information. the hardest to automate. ANI requires ‘thinking’ but has failed is already advanced enough to to do most of what people and Until AGI or near-AGI is achieved handle a wide spectrum of jobs, animals do ‘without thinking.’” this adaptation of environments from creating nutritious vegan will remain a necessity. But as “meat” to predicting the success Until AGI is achieved, ANI is is evident in McCann Japan’s of a startup. reliant on machine learning recent appointment of an AI algorithms which require huge as Creative Director, this gap For approximately the next 25 data sets and an adaptation of between ANI and AGI is closing years, ANI innovations will be our natural world to be capable fast. restricted to carrying out specific of comprehension. In the words CHRIS KREINCZES tasks, or a defined group of of John Martinis, a quantum EDITOR & CREATIVE specific tasks. Processing physicist currently working at DIRECTOR, SPRINGWISE jobs considered complex Google, “machine-learning to the human brain — such algorithms are really kind of as advanced mathematical stupid. They need so many calculations — will be easy work examples to learn.” The ANI for an ANI. Conversely, it’s often examples below are only able the most mundane tasks — for to perform their specific tasks example, differentiating between because humans have already 4 Powered byAt a Glance WHAT AIME use big data and AI algorithms to pinpoint the geolocation of dengue fever outbreaks, three months before WHO AIME WHERE United States CONTACT www.aime.life infoaime.life AI can predict where epidemics will happen AIME use big data and AI algorithms to pinpoint the geolocation of dengue fever outbreaks, three months before they occur. We’ve seen how speed and direction, rain concern to public safety, autonomous drones could volume, housing types, and such as riots? be used to protect human more. It then calculates a populations from mosquito- percentage of outbreak borne outbreaks, and now probability in a given area. artificial intelligence is The platform determines predicting where human the epicenter of a epidemics can occur. AIME spreading disease within (Artificial Intelligence in a 400 meter radius, and Medical Epidemiology) this will be the crucial area has developed a Dengue that health officials look to Outbreak Prediction contain and monitor. platform that predicts outbreaks of the potentially Having successfully fatal disease three months trialled their platform with in advance. 87 percent accuracy in dengue fever outbreaks, The AI algorithm combines the team hopes to target epidemiological research other highly contagious with other data known diseases, such as Ebola to impact the spread of or the Zika virus. Could insect-borne disease, such AI algorithms be used to as population density, wind predict other scenarios of 5 Powered byAt a Glance WHAT Using AI to find audiences for new content, smart video tech company Unreel Entertainment launches Unreel.me, a platform WHO Unreel Entertainment WHERE United States CONTACT www.unreel.me hellounreel.me Tool for build-your-own video streaming sites Using AI to find audiences for new content, smart video tech company Unreel Entertainment launches Unreel.me, a platform for custom-built streaming sites. Unreel Entertainment uses automated highlight reels. Unreel, with content owners AI bot Bumblebee to trawl Once an Unreel.me site is retaining 85 percent of their the internet and social media live, the Bumblebee bot earnings. for the most popular videos, begins including the new making it easier for viewers content in its analysis and Finding ways to reach to find and share new suggestions to viewers. multiple audiences is a content. Now, with Unreel. Content owners can choose constant challenge, and a me, everything a user needs to monetize their site number of platforms are to create a bespoke video and create paywalls and helping content creators streaming site and app, subscription plans. Unreel.me aggregate the management and build an audience, is also help creators advertise of their output. We’ve seen provided in one place. and sell merchandise. an online tool for building apps and widgets for multiple Each Unreel.me site and As part of the Unreel.me devices and a platform app uses content uploaded launch, Unreel Entertainment that helps video producers directly to the Unreel has raised financing for manage content and engage platform, and syncs with a USD 1 million Creator socially all in one place. videos posted on social Appreciation Fund. The Fund networks. The platform is offering owners of each What other tools could help provides everything from Unreel.me site 100 percent of content providers manage an easy-to-use content all earnings until the money audience and output? management system to is dispersed. After that, any time-stamped commenting revenue generated will be and sharing capabilities and split between creators and 6 Powered byAt a Glance WHAT Ross Intelligence, which is built on IBM's super-computer Watson, uses natural language processing to answer legal questions. WHO Ross Intelligence WHERE United States CONTACT www.rossintelligence.com teamrossintelligence.com AI lawyer speeds up legal research Ross Intelligence, which is built on IBM's super-computer Watson, uses natural language processing to answer legal questions. Lawyers have to maintain program reads through and recall vast amounts the entire body of law and of information in the form returns a cited answer of legislation, case law as well topical readings. and secondary cases, Ross also monitors the law and they spend up to a constantly to keep the user fifth of their time on legal updated about changes research. But an AI app that might affect their case, called Ross Intelligence so they don't need to sift could soon help with that. through the mass of legal The program, which is built news. on IBM's super-computer Watson, uses natural Ross is already being language processing to used by law firm Baker answer legal questions and Hostetler, who will in a fraction of the time be applying it to their that it would take a legal bankruptcy practice. assistant. Could a similar program be created for the medical To begin, legal profession? professionals can ask Ross a question as they would ask a colleague. Then the 7 Powered byAt a Glance WHAT Algorithmic logo-makers Tailor Brands will now generate unlimited branding material for companies via a subscription WHO Tailor Brands WHERE United States CONTACT www.tailorbrands.com contacttailorbrands.com AI design service can automate a full branding package Algorithmic logo-makers Tailor Brands will now generate unlimited branding material for companies via a subscription service. Good design and visual provides companies branding is no longer the with unlimited branding privilege of companies packages on a subscription with the money to pay basis. great in-house designers or outsource the work. We Tailor Brands Studio will have already seen startups harness data from the 13.5 thinking outside the box to million designs created by provide more affordable the company's algorithms design services. This in the past two years and platform offers unlimited, use it to offer automated small graphic design tasks designs of logos, business to other startups through cards, printed stationary a subscription package, and social assets. and Tailor Brands uses Companies can sign artificial intelligence to help up for the service on a business automatically subscription basis for USD generate a logo that fits 9.99 a month, which gives their company. Now, Tailor them unlimited branding is expanding its tools by material. The company will launching Tailor Brands add merchandise and ad Studio — an algorithmic, design later this year. automated service that 8 Powered byAt a Glance WHAT Player XP uses machine learning to help identify constructive feedback in mobile video game reviews. WHO Player XP WHERE United Kingdom CONTACT www.playerxp.io helloplayerxp.co.uk Game feedback platform uses machine learning to filter reviews Player XP uses machine learning to help identify constructive feedback in mobile video game reviews. Sifting through reviews to categorizes the data via find out how to improve date published or ratings, the video games they are so more in-depth analysis building can be a tedious can be done. task for developers, but a new program The platform is dedicated for aggregating game to mobile games, and feedback could help. currently aggregates more Player XP is a review than 40 million iOS reviews and analysis platform for from 250,000 iOS mobile mobile gaming. games. Could machine learning be used to help The aggregator uses players themselves find machine learning new and exciting games and natural language that fit their tastes? processing to analyze the reviews. It filters unhelpful reviews or spam, showing only constructive feedback through its artificial intelligence program. As well as removing unhelpful feedback, the platform 9 Powered byAt a Glance WHAT Pefin is an artificial intelligence financial advisor that monitors over two million data points to provide decisions. WHO Pefin WHERE United States CONTACT www.pefin.com hellopefin.com AI financial advisor makes 80 year projection Pefin is an artificial intelligence financial advisor that monitors over two million data points to provide decisions. Being well-informed and begin, customers set up How else could AI be used able to view things in the an account and connect to help people plan better longterm can make all the it to their other financial for the future? difference when it comes accounts, creating a to financial decisions. We financial profile. They then have already seen the likes input their long term plans of a tech assistant using — such as having children, big data to help people get buying property or going the best exchange rates, on holiday. These can be and now Pefin can help changed at any time. Pefin even more. The artificial then takes into account intelligence platform, inflation, taxes, market which is currently in Beta, conditions, government is a financial advisor that is rules, and even changing significantly cheaper than daycare costs, and creates its human counterparts. an 80 year financial projection. Then, whenever Pefin uses artificial neural anything changes, either networks to constantly within the customer’s monitor over two million personal finances or in data points, and provide the global economy, Pefin data-driven decisions updates the profile and via machine learning. To projection automatically. 10 Powered byAt a Glance WHAT Chile-based NotCo replicates the taste and texture of animal food products using AI, biochemistry and plants. WHO NotCo WHERE Chile CONTACT www.thenotcompany.com whynotthenotcompany.com Artificial intelligence uses algorithms to make nutritious vegan meat Chile-based NotCo replicates the taste and texture of animal food products using AI, biochemistry and plants. Situated at the intersection of tries to enhance nutrition food tech has focused either food and technology, NotCo by including additional on the processes of buying, (The Not Company) used sustainable ingredients. selling and delivery, such as artificial intelligence to create Products available so far are paying for coffee with a smart the world’s smartest chef. NotMilk (made from nuts, rice, cup, or finding new methods Nicknamed Guiseppe, the peas, linseed, and coconut), of sustainable production advanced algorithm examines NotHotdogs, NotChocolate, — even if it is in the form of the molecular structures of NotMayo, with more in desktop hive farms. NotCo meat and dairy products and development. takes the tech directly into attempts to replicate them the kitchen, creating recipes using plant material. With animal agriculture designed to be used by the using nearly a quarter of the average cook. The company says the world’s water supply and 45 secret to creating notably percent of available land, Are there other industries good-tasting vegan meats, meat and dairy products are that could use plant-based, eggs, milk and cheeses is to especially unsustainable and molecular-level replication to find plants that have similar environmentally unfriendly. increase sustainability? molecular structures. The By helping people eat final product should have more healthily without almost identical flavors and compromising flavor and textures. When Guiseppe texture, NotCo hopes to lead creates a recipe, the a revolution in sustainable algorithm not only considers protein. the relative availability of various plants, it also So far, much of the work in 11 Powered byAt a Glance WHAT Autism Glass is a wearable, behavioral aid that uses machine learning and real-time social cues to help those with autism interact WHO Stanford University WHERE United States CONTACT www.autismglass.stanford.edu autismglassstanford.edu AI glasses helps children with autism read facial expressions Autism Glass is a wearable, behavioral aid that uses machine learning and real-time social cues to help those with autism interact with others. In the US, over one To use, patients put on understand their progress million children suffer the wearable glasses, and even watch back their from autism. These which incorporate an interactions. children find it difficult outward-facing camera. to recognize emotions Then, the system uses Research is still undergoing through facial expressions, machine learning and after a successful 40 making social interactions artificial intelligence to person pilot study, and the very challenging. While understand the facial developers are currently patients can gain an expressions of people seeking participants on understanding through the wearer encounters, the autism spectrum aged behavioral therapy, this can and provides cues 6-16. We have already be both time-consuming about their companion's seen an app that gives and expensive. Now, emotions in real-time. users friendship advice Autism Glass is a wearable The system records the by monitoring their heart aid from researchers amount of eye contact that rate. Are there other new at Stanford University, takes place, which can technologies that could that uses Google Glass, provide insights into the help autism patients better machine learning and real- wearer's interactions. The understand the world? time social cues to provide wearable is connected to those on the autism a smartphone app so that spectrum with another the patient, their parents option. and medical professionals can monitor and 12 Powered byAt a Glance WHAT AI-CD β is a robot that is being treated as an actual employee in an ad agency. WHO McCann Japan WHERE Japan CONTACT www.mccann.co.jp miyoko.ohkimccannwg.com In Japan, an artificial intelligence has been appointed creative director AI-CD β is a robot that is being treated as an actual employee in an ad agency. Advertising and media and messages. entertainment, including are often at the forefront music, movies, and TV of new technology, The AI-CD β can work drama, so we’re very and we have already through far more enthusiastic about the seen augmented reality historical adverts and potential of AI-CD β for platforms showing content commercials than any the future of ad creation. in the real world and a human advertising The whole company is virtual reality advertising creative, and can therefore 100 percent on board to network for brands. aid its colleagues in the support the development development of ads for of our AI employee.” Now an artificial clients. But the robot is intelligence robot, AI-CD β, also being treated as part How else could artificial developed by Japanese of the team at McCann, employees be used to advertising and marketing taking the title of "creative complement, rather than agency McCann Japan, is director" and attending the compete, with human co- set to work on providing opening ceremony for new workers? new creative direction for company employees. commercials. The AI will give input on projects, McCann Japan CEO mining and analyzing Yasuyuki Katagi said: creative databases of “Artificial intelligence is adverts to find the best already being used to commercials for products create a wide variety of 13 Powered byAt a Glance WHAT Moodbox is an emotionally intelligent speaker system that uses artificial intelligence to learn to pick the best music. WHO Ivo WHERE Hong Kong CONTACT www.mymoodbox.com infoivo.hk Smart speaker reads emotions to play music Moodbox is an emotionally intelligent speaker system that uses artificial intelligence to learn to pick the best music. We have already seen or time of day, Moodbox fully voice-controlled, and a headphones that stimulate can make music choices, number of Moodboxes can nerves to let users feel reacting to, for example, be set up in a house and their music. Now, a new how the user feels first linked together, playing speaker system is claiming thing in the morning, or music at the same time. to use human moods to when they are coming learn what tunes users home from work. How else can artificial want to listen to. intelligence be used to Hong Kong based- improve the way we listen Moodbox uses its artificial developers Ivo say to and discover new intelligence system to work Moodbox uses an music? out what music users most emotional intelligence want to listen to. Users program known as Emi. interact with the Moodbox Emi learns different moods   by voice control, and and plays music and can keep a diary of their lighting accordingly. The moods using the speaker's program uses AI to analyze smartphone app. The the voice of the user and system can work out what their musical preference, music to play by listening then picks out the perfect to what users say and tracks from millions of their tone of voice. Then, songs and lyrics. The depending on the situation speakers are wireless and 14 Powered byAt a Glance WHAT Webpage.ly uses an algorithm that suggests the most searchable keywords, helping developers optimize their sites for better WHO Webpage.ly WHERE Canada CONTACT www.webpage.ly helpwebpage.ly SEO artificial intelligence could boost webpage ranking Webpage.ly uses an algorithm that suggests the most searchable keywords, helping developers optimize their sites for better ranking. As search engine rankings page rankings to suggest become evermore keywords that enable important, companies developers to produce are hiring costly SEO higher-impact SEO specialists to help their content. While suggestions sites gain more exposure are made by the artificial and click-throughs. But intelligence, users will for smaller startups that still have the final say by cannot afford to hire approving content for their someone, Webpage. website. ly could offer an affordable alternative. The We have already seen AI subscription platform aims software being used in to help web developers the workplace for a variety make their sites more of functions, from writing searchable. personalized emails to hiring the best team fit. Webpage.ly uses an How else could artificial algorithm that learns intelligence optimize how to boost contextual workflow? webpage content over time. It analyzes visitor search behavior and 15 Powered byAt a Glance WHAT Val.ai is an artificially intelligent search engine for films, which can process natural language and analyze thousands of variables. WHO Valossa WHERE Finland CONTACT valossa.com infovalossa.com Filmbuff search engine helps users find movies they like Val.ai is an artificially intelligent search engine for films, which can process natural language and analyze thousands of variables. Finding new video content company claiming their online, or searching for software as "the most films without knowing advanced video analysis the title can still be a platform". challenge, even with the most up-to-date search The platform can examine engines. That's why Val. advanced questions ai, a new video search and queries to find films. engine from Finnish Users can type in, for company Valossa, is using example, "show me sci- an artificial intelligence- fi movies about space powered tool that can battles and laser guns", analyze thousands of or "find me epic history variables in movies and movies". The platform also videos, and give more features a descriptive accurate searches. voice-controlled search engine for discovering Val.ai is capable of entertainment platforms. analyzing video streams and identifying thousands How else can search of themes — the artificial engines be improved to intelligence looks deep make smarter and more into the content, with the useful searches? 16 Powered byAt a Glance WHAT Smart Moderation uses artificial intelligence to block spam messages and comments on social media. WHO Smart Moderation WHERE United States CONTACT www.smartmoderation.com infosmartmoderation.com Artificial intelligence learns to block spam messages and comments Smart Moderation uses artificial intelligence to block spam messages and comments on social media. Moderating social media reviews like a human. The digital profiles? feeds or online comments dashboard lets users view is important for marketers all comments from different and entrepreneurs, but social pages, and can be looking after a business’s set to automatically hide social media accounts inappropriate comments. can be incredibly time Users have the option consuming. Now Smart to receive daily reports Moderation is bringing of spam comments, and intelligent automated can also switch to manual moderation to Facebook moderation. The platform and Instagram profiles. performs more traditional spam-blocking tasks too, The program uses an AI such as clearing adverts. software that understands natural language. It We have seen artificial processes the meaning of intelligence being used sentences, rather than just to turn spreadsheets into flags for keywords, and written reports and clean removes trolling or abusive up inboxes. How else can comments accordingly. smart online programs The company says the streamline the process software learns and of managing businesses' 17 Powered byAt a Glance WHAT Test4startup uses an algorithm to test business idea strength in current markets and against competitors, generating feedback WHO Test4startup WHERE Russia CONTACT www.test4startup.com supporttest4startup.com AI algorithm predicts startup success Test4startup uses an algorithm to test business idea strength in current markets and against competitors, generating feedback and advice. We’ve seen how algorithms recommendations based are being used to help on how a product or build strong startup teams, service compares to others and now Test4startup can — recommending a price help founders test the reduction, for example. 'successability' of their Startups can test their actual idea using artificial products and business intelligence software. ideas from USD 10, or for USD 15, they can also Users begin by receive feedback from uploading an overview focus group professionals. of their business aims. Test4startups How else could algorithms then implements a be used to analyse context keyword-based analysis and provide feedback? that aggregates information relative to the specific market, providing infographics of competition, trends and pricing variation. Test4startups algorithm is then able to provide 18 Powered byAt a Glance WHAT The Deep Neural Network software helps quadcopter drones identify trails and paths, and lead them to lost or injured hikers. WHO University of Zurich WHERE Switzerland CONTACT www.uzh.ch sdavideifi.uzh.ch Drones better than human rescuers at following mountain pathways The Deep Neural Network software helps quadcopter drones identify trails and paths, and lead them to lost or injured hikers. Every year in Switzerland, uses the algorithm to learn areas to minimize risk to emergency centers trails and paths through human workers, and its respond to around 1,000 a pair of small cameras, AI can identify paths and call outs for lost and interpreting the images avoid crashing without any injured hikers. It can often and recognizing man-made human involvement. take hours and significant pathways. Even when manpower to locate lost working on a previously However, the developers mountaineers, but new unseen trail, it was able to say the technology is still software for quadcopter guess the correct direction in its early stages, with new drones is making the hunt in 85 percent of the cases. software R&D required to quicker and easier, and has The drones' speed and make the drones able to the potential to help find accuracy make them more identify human targets and human survivors in disaster effective than human relay this back to rescue zones around the world. trackers. teams. The drone uses a The researchers hope that We've already seen computer algorithm eventually multiple small ambulance drones helping called a Deep Neural drones could be combined citizens give lifesaving Network. The program was with human search and procedures before developed by researchers rescue missions, to cover professionals arrive. How at the University of Zurich more terrain and find else can drones and robots and the Dalle Molle people faster. The drones help rescue workers save Institute for Artificial can cover terrain quickly more lives? Intelligence. The drone and check hazardous 19 Powered byAt a Glance WHAT MurdochBot uses IBM's sentiment algorithm, Sky TV viewer data and users' conversations to offer entertainment recommendations. WHO MurdochBot WHERE United Kingdom CONTACT www.talkingquickly.co.uk bentalkingquickly.co.uk Bot suggests TV shows based on group chats MurdochBot uses IBM's sentiment algorithm, Sky TV viewer data and users' conversations to offer entertainment recommendations. Instant messaging app questions about their eclipse social media by Telegram is growing in favorite TV shows. The bot offering more personal popularity with its multi- then monitors the content connectedness, a new platform integration to understand not only space is created for data- and encrypt-able P2P what kind of shows users mining and delivering messaging. Telegram will like, but also in what content. How could programming has an open context and times they companies make the API so users can add bots would like to hear about most of this opportunity, into chats, and one such them, learning more as while providing users feature is MurdochBot, conversations develop. with autonomy over their which hopes to suggest TV Founder Ben Dixon privacy? shows users will like. believes that analyzing natural conversations Murdochbot is an AI will give businesses a protocol that uses better idea as to what sentiment-analyzing consumers want. Unveiled algorithms developed at the TechCrunch Disrupt by IBM, in combination London Hackathon this with Sky TV’s archive of year, Murdochbot has yet viewer data. Users first to reveal development add MurdochBot into plans. any conversation, and it will ask them a few basic As instant messaging 20 Powered by