england and north wales. it will peter out overnight, and will leave clear skies in the north ahead of this next weather system. but much colder more widely overnight, and a more widespread frost. but again a lot of cloud around. and this becomes instrumental, this weather system, later in the day tomorrow when it brings snow to the hills across many northern areas and even a smattering further south. it is still cold air. let's look at tomorrow evening. a band of rain turning readily to snow, particularly over the hills. behind it, more cold air and mmore wintry showers. and more wintry showers. hello this is bbc news. the headlines. theresa may will brief ministers on the brexit discussions she's been holding with party leaders and senior mps. police speak to the duke of edinburgh who was seen driving without a seatbelt — 48 hours after being involved in a crash near sandringham.

president trump's latest offer to end the longest government shutdown in us history is rejected by democrats. two people have been arrested following a car bomb attack in londonderry last night. two people have been killed and 1a others hurt in a fire at the french ski resort of courchevel. defending champion roger federer is knocked out of the austrialian open in the fourth round. now on bbc news, it's time for click. this week, we talk talk. but can your voice reveal what's wrong with you? can this helmet hear you? and would you all please stop shouting? alexa, what are the amazon spheres?

they are an unusual place to work, i have to say. the domes house 40,000 plants. they also have waterfalls, meeting spaces and shops. they're slightly tropical, quite moist, and give you a real sense that you're working in the future. i mean, who'd have thought in the middle of seattle, there'd be a rainforest? an amazon rainforest? say "amazon", think what? jungle? 0nline shopping? or do you think of alexa?

this is, after all, the company that has brought voice control to the masses. there's no training required for using voice. the vision for voice is that anyone — young, old, in any country, in any context — can use natural language to interact with technology. alexa is only gonna be as successful as the partners are that support it. and that's why voice is the next battleground for the tech supergiants. the more voice apps there are and the more comfortable we are talking to them, the more they will integrate themselves into our lives. just this week, in fact, alexa‘s voice has started,- sounding more human. developers have used artificial intellidence technioues to learn _ from news broadcasts how to intone and emphasise words in a sentence. now, as a human being,

you'll know that you can tell quite a lot from someone‘s voice — not just what they mean but also how they are feeling and possibly whether they're ill. now, that's obviously really hard for a computer to do, but nick kwek has been looking at some technology that is working towards just that. i've come to tel aviv, israel, to see a doctor with a difference. no cold stethoscope, but i do need to say ahh, as well as make a bunch of other sounds. it's a early decay of the voice. today, dr levanon is going to give me my medical, purely by listening to the way i talk. ..see anything beyond verbal of what my state of health is. he's a firm believer that it's not what you say, it's how you say it. je t'aime. wo ai ni in chinese. or in hebrew, ani ohev otah.

it's — all of that are ways of expressing love. and if i lack that and i commanding you, it's the same in all the languages. he isn't a doctor of medicine, but of research he's been working on for more than 2a years, which has culminated in a new piece of software. it's a tuesday morning. we've got to catch a flight later back to london. by analysing my vocal intonation, dr levanon says his algorithm can infer my mood, attitude and personality. you are ambitious, you are curious, a little bit adventurous, and a little bit conservative. it is something in between. ifeel like i'm getting my palm read by a psychic. 0k! and the graph is saying, "0h, you are very tired." i am very tired!

but the reason i am here is for something far more important, because dr levanon says his system can also be used to diagnose whether i have a disease, and even how bad it is, just by listening to my voice. we can show the early signs of parkinson's, heart problem, autistic children, early signs of prostate cancer. he's comparing the audio wave patterns of someone‘s voice against an archive of recordings of people with known illnesses. how are you analysing the wave patterns? so you can see that there is a weakness in this area of frequencies. you can see the same crater but also vibration here. these are the signs of ill person and you can see decay of the voice, the deterioration — is ill, tired, can't express himself well.

if you are combine all the signs together, we can see how severe is problem and what is his health problem. apparently, the system is 75% accurate and can tell you the severity of your condition. if for real, integrating the tech into listening devices could change the way we interact with our gadgets. further down the road, with our technology, these machines would also be able to tell "hey, something is different with your health condition. maybe you want to go and do another check—up." maybe it will let my doctor know about this. so from the voice, there is so much information that can be gained, and we are just at the beginning. 0k, reality check time, because as amazing as this sounds, you might be thinking this is all too good to be true. but the thing is voice is already used to diagnose certain conditions and medical professionals do see

that this could have more potential. in london, i went to meet dr nijjer, a consultant cardiologist at hammersmith hospital, and honorary senior lecturer at imperial college london. well, there's no doubt that doctors use a patient‘s voice as a way of diagnosing certain conditions and there are neurological conditions like motor neurone disease or parkinson's that have characteristic changes to the voice to help us make a diagnosis. but whatever a voice‘s potential, today beyond verbal has had only one paper published, in conjunction with the us—based mayo clinic, investigating the connection between voice and coronary artery disease. what this study has done is attempted to make a diagnosis of coronary artery disease using the voice alone. this is a completely new avenue and i would say that the technology at present needs a lot more work. the study that's been published doesn't give us definitive proof that this technology can make that diagnosis. the study has been performed in a highly enriched and highly selected group of patients.

and actually, i would say the capacity of voice to be able to detect the severity of an individual lesion, as they've proposed, is a little far—fetched. california—based telemedicine start—up salubermd has begun incorporating beyond verbal‘s mood—analysing system into its gp appointment app. 0ne second. bear with me while i bring up your data here. whilst remote doctors listen to problems over video call, the software can inform them about their patient‘s emotional state. how accurate is beyond verbal around the mood? very. how is accurate regarding the acute coronary syndrome? we don't know yet. at present, there is a very good feel around it but i, as a doctor, cannot trust a feel. i need to trust real correlation or real demonstration. beyond verbal isn't the only business claiming to understand our health via our voice.

healthymize, sonde health, peakprofiling and even ibm are just however, it's still not enough to say "that's it!" it's still enough to say alert, be — take care of yourself, go to the doctor, make other checks. could come with repercussions. and do we want the tech giants to know so much about us? the problem with anything that's predicting a disease process is that it can be wrong or its sensitivity is not high enough or it's not specific. so it will often make diagnoses when there isn't that particular problem or it'll over—diagnose problems that don't need any investigation, so that would be a concern to me as a doctor that we end up testing too many people. and it would also be a concern for patients,

because they may end up being told they have a problem when actually, there isn't that concern. further trials are being conducted in israel, china and soon, the uk. and whilst it might be simply sci—fi, the idea of alexa in the future alerting me about my health is at once appealing, yet unsettling. hello and welcome to the week in tech. this was the week that the world came together for a truly noble cause. yes, it's my great honour to officially confirm that this stock photo of an egg is now the most liked instagram photo ever. truly inspiring! in non—egg news, 0range shirt kid joined the list of people accusing video game fortnite of allegedly stealing dance moves. and the chinese space agency shared images of the first seed to sprout on the moon. youtube has said it will now ban clips of dangerous and harmful pranks from the platform. the change comes in response to so—called challenge videos which, in some cases, have resulted

in injuries or even death. following through with the change may be difficult, though, with critics arguing it could be hard to decide where to draw the line on what constitutes ‘dangerous‘. time to make room on the bedside table for yet another essential gadget. sportswear company nike has unveiled the latest incarnation of its self—lacing smart shoe. controlled by an app, the $350 trainers adapt to your foot with a range of lace tensions. and if you're worried about getting stuck in them, no fear — the company says the battery will last at least ten days. and finally, spare a thought for the robots of henn na hotel this week, some of whom are looking for work. in a twist to the classic tale, japan's iconic robot hotel has had to lay off half of its 243—strong robotic workforce. velociraptor receptionist is gone, as well as dozens of the in—room assistant churi. as well as maintenance issues, churi kept on waking people up at night, mistaking snores for questions, and velocira ptor receptionist couldn't handle foreign guests. and now, we're going

to go back in time to the consumer electronics show in las vegas from last week, where voice was a massive deal. and we asked lara lewington and richard taylor to check out the big talk at the show. it's only taken a few years for voice—activated smart assistants to become an integral part of every self—respecting techie‘s lifestyle. amazon's alexa started out as a small, self—contained box four years ago. today, it's enjoyed a phoenix—like rise, compatible with over 28,000 smart devices and myriad gadgets of all shapes and sizes. alexa, run the video. this bot vector isn'tjust cute, it can also order you a pizza. fire up this alexa oven and pair your smart—cooked breakfast with a smart—brewed coffee. and when you're feeling the creative juices flowing, you can always kick back with your smart instrument. if you're a keyboard player, you know it can be really

frustrating navigating your way through menus, just to get to what you want. but with alexa built in, there's a world of possibilities that open up. so, for example, i can say "alexa, ask go:piano to select strings". alexa: the tone is tuned to strings. and with this particular device, there's a number of built—in commands that work offline. so, for example, "head up display". and there we go! ready to roll! google assistant may have been late to the party but is trying to catch up. it is built into this alarm clock with internet phone... hey, google. what's the weather? ..or is around for a chat whilst you get ready. google assistant: currently in las vegas, it's 57 and partly cloudy. meanwhile, apple — and to a lesser extent, microsoft — are making a play

for voice domination too, but amazon and google remain clear front—runners. not that you need to commit to just one. with this bluetooth headset from 0v loop, you can simply say "0v phone" to activate either siri or google, and "0v alexa" to activate its amazon rival. play gorillaz. but the voice—activated assistants are hoping to break new ground in specific environments. chris is designed to work when you're in the car, even when you're off—line, letting you control music and even navigation with your voice and, somewhat curiously, with a few gestures too. and back in the home, some specialist devices are aiming to use voice function for slightly more human purposes. this is a social companion robot for the elderly. it combines google's voice recognition technology and niq's it combines google's voice recognition technology and leq's

own platform to be able to not just respond but also initiate conversation. it can actually provide notjust reminders but also goals for its users. so if somebody has been sitting down for too long, the motion sensors will alert and it may get them to get up and do a bit of exercise. it could also be helping them keep in contact with their family through the tablet elements. but of course, these devices do have a way to go before they provide a totally seamless experience. and now it is going to pull out one of the... voice should be the most natural user interface that is out there, versus touch or using a mouse or keyboard, or anything else, and so making it conversational, like you and ijust having a conversation right now, that is how it should be with alexa, right? we shouldn't have to keep using her name, we should have ideas of knowing what we are about to say. this is a gesture controlled

speaker, which works like this. the company behind it are hoping to install their technology in virtual assistants, and it is notjust about being able to wake you up by tapping, it is also going to track how far away you are in the room so that the volume can be adjusted accordingly. a device's built—in speaker emits ultrasound waves and its microphone receives them, mapping where you are. the software acts on this information, it may be pausing something you're watching or listening to, or turning off heating and lighting. while there are ideas aplenty about how the future of voice could play out, for this year's ces, google and amazon certainly seem to be enjoying having the last word. thank you very much, goodnight. in the future, we will talk to our devices, and they will talk back.

and that means that developers need to think in a much more nonvisual way. when you glance at a screen, you can selectively absorb a lot of information and selectively pick out what's interesting and relevant to you, but when you're listening to a skill or voice application in action, you're forced to listen to whatever the skill is telling you. so it needs to get it right and it needs to get it right every single time. jargon was one of nine start—ups selected for amazon's 13 week alexa accelerator programme, which helps develop ideas that use voice and send them on their way towards more funding. despite having more than 10,000 alexa developers working inside amazon, the company is also inviting in ideas from start—ups, from simple adaptations of home appliance controls to work in an office environment to much more complex speech recognition tasks, which is working towards understanding people

who have problems with their speech. i love you. a small company comes up with a great idea that inspired us and we said, wow, we could really help this company by creating new apis that they could take advantage of. a company came in with our first accelerator, and they have this really interesting service where you read a physical book to your kid typically, and their backend is listening to your reading, knows where you are in the book, and at certain points in time will play a sound effect. are you afraid of monsters? so imagine you are reading a harry potter book, and at the right point,

there are storm effects that come on, right? and it's a magical experience. and yet, we're still only in the early days of speech recognition and there are plenty of hurdles still to overcome. we have made huge advances in voice control in a really short time, butjust where are we on the speech recognition journey? well, that's what kate russell has been looking at. what does the internet sound like? dialup internet dial sound playing. no, no, that was 20 years ago. what does it sound like today? i'm siri. i'm alexa. jumble of computerised voices plays. but as we saw from ces floor earlier, the age of the voice assistant is well and truly upon us. around one in ten uk homes already have a smart speaker, and if you're one of them, you might want to mute its microphone for the next four minutes. more people now have mobile phones than have toilets, and of those people,

60% are already using voice to talk to their phone. experts predict that by next year, a third of all web browsing could be done without a screen. in a world where 66% of us say they want their lives to be simpler, anything that removes friction, just by removing buttons and just letting us talk to things, is going be attractive. but a post browser world raises an interesting question for all kinds of businesses. how will our customers find us? it is a head scratcher for broadcasters, like the bbc, because without an audience, we might as well pack up and go home. so voice platforms are both really, really exciting and really, really challenging. they're exciting because they are an infectious way to get audiences to your content. it is also a challenge though because you have other companies, amazons, apples, googles coming in,

saying what does this mean for other businesses? and we, as the broadcasters, need to think really, really carefully about how we work with those partners in order to make sure that we protect our business an do the right thing for audiences, and don't end up killing radio by accident. there is also the issue of search and discovery. for audiences to find content, it needs to rank highly in search results. if you type ‘brexit‘ into a browser, you get a page of links to choose from. when you say to alexa "tell me about brexit", she gets to choose — or more importantly — amazon: getszto choose whichztop ; to you. which. new based on interests you have already experienced. so how does click get to the top of google's list? it's not our list, it's the user's list, that is what is important.

we are simply the distribution of the content that already has a huge audience, so there are certain players that you just have to have, and there are other times when there is just unique experiences out there that a partner has, that we believe is what the users really want. recipes is another, we said we have to make sure that we have the right recipe partners because that's a key use case for speakers. those are the two axes that i would say really ranked on what the users really want, and also the category and the popularity of a lot of content that they already have, which has already been established before we even start to think about the partnership. so you are going to see the same kind of stuff over and over, but what if you want to watch something other than bbc click? the serendipity of discovery is already under siege by recommendation engine code. do we really want to get caught up in a filter bubble

of just one result? the platform owners really — they know they need to work on discoverability. if you have either an echo show or a google home hub which has screens on, you will see that they are trying to suggest things you need to have all the time, and you can also look at the skills that you have enabled. useful. anyone who has a smart speaker will know how hard it is to remember all those skills and actions, and exactly what you have to say to use them. it can feel more like speaking to a truculent two—year—old than an intelligent machine. but there are opportunities too. we are experimenting as a research company with using speakers to ask you questions, so in the morning alexa might say to you, "hello, would you like to answer some questions about breakfast cereals today? i will give you a £20 amazon voucher." and that might be much easier than me sending an e—mail and you clicking on it and answering

some questions on your phone. some of you will be horrified by the prospect of a world full of talking technology. we all saw what happened with hal, right? sorry, i don't know that one. a likely story. but for most of us, voice promises to improve our relationship with technology — eventually. and that is it from seattle, at least for now. don't forget you can follow us on social media, facebook, youtube, instagram and twitter. we're @bbcclick. thanks for watching and we will see you soon. hello, another cold day awaits.

it was bitter this morning. the cold theme stays with us for much of the week. at times, it'll feel colder still, particularly when the wind picks up. the biggest risk of snow is monday night. there will be widespread frost. hopefully, we will see more sunshine for the coming week. we have seen sunshine today across east anglia compared with yesterday, coming through behind the weather front. by the end of the day, it will come through more readily across scotland and northern ireland. we have showers for western areas and a weakening weather front giving rain and hill snow. still cold air out and about. this evening, we keep clearer skies in the south and east. it will turn cold pretty quickly across scotland and northern ireland. behind the weather front,

temperatures down to several degrees below freezing. we already have more cloud piling in ahead of our next weather front. that high cloud will obscure the view of the lunar eclipse. this is quite a main player in the weather story. how much snow it will give and at how low levels. starting quiet but cold on monday, mist and freezing fog. this looks more active as the winds strengthen bringing rain into northern ireland and scotland in the afternoon, turning to snow over the hills. through the evening, you can see a risk of snow far south, even the hills could see a smattering for a time. it will clear. more likely to settle on the hills further north with several centimetres. showers will follow.

with the cold air behind, temperatures on tuesday evening in the sunshine just 5 degrees. —— even in the sunshine. with lots of sleet and snow showers following. staying cold. there remains the risk of further light snow. looking dry on wednesday, thursday, it is still pretty cold. as ever, there is more on the website. good afternoon. theresa may will talk with her cabinet later before addressing parliament tomorrow on how she intends to break the brexit deadlock. meanwhile, with britain due to leave the eu on march the 29th, at least one cross party group of mps are discussing how parliament could take control of the process to avoid a ‘no deal‘ brexit. 0ur political correspondent nick eardley has the latest. the clock really is ticking. 0n the

29th of march as things stand we leave the european union. the government struggling to find a