1:30 pm

this is bbc news, the headlines. two british nationals have been arrested on suspicion of attempted murder after a mass stabbing on a train from doncaster to london. five people remain in hospital, two with life-threatening injuries. police have declared a major incident and say they are still working to establish the motives for the attack. at this stage there is nothing to suggest that this is a terrorist incident. the bbc speaks to eyewitnesses, who describe the terrifying moments they experienced whilst on board the train.

1:31 pm

i looked at my hand and it’s covered in blood. then i looked at the chair and there’s blood all over the chair. then i looked ahead and there’s blood all over the chairs. so i’m thinking, ok, this is pretty serious. with ai, searching through this content gives you the ability to understand the context and the emotion behind things. so you could say, how did this person feel?

1:32 pm

the timeline for a fully autonomous airport is set on 2050. i’m very surprised in the last two years how far technology has matured. that’s crazy, that something so unassuming can be so transformative. this got us one of our ten nobel prizes. newsreel: the battle of britain is on. goering is pouring his air force across in waves to try and break our spirit. it’s been 85 years since the battle of britain, an aerial campaign fought during world war ii - and today, i’ve come to the imperial war museum in london to learn more about it and other conflicts. but i’m not seeing the latest exhibition. i’m here to listen to survivor stories being brought to life thanks to new technology. they’ve digitised almost 46,000 recordings from their archives. and now, with the help of ai, they’re becoming more searchable than ever before.

1:33 pm

one month we did 70 hours flying each. they’re really far ranging. often they’re three hours long or more, and we discover an awful lot about their experience of war, both on the front line but also at home, and some incredible stories there. in collaboration with google cloud and capgemini, the museum has fed almost two thirds of its audio collection through an algorithm, creating an ai powered catalogue that acts like a human archive assistant, but with superpowers. over 8,000 interviews, over 20,000 hours of audio. it would take someone over 20 years to transcribe the recordings, but the ai has done it in a few weeks - and this new interactive platform aims to make them more accessible. with ai, searching through this content gives you the ability to understand the context and the emotion behind things. so rather than saying we’re looking for this date

1:34 pm

or this person, you could say, “how did this person feel about ve day? “were there any funny stories that happened “whilst you were in africa?” and these are questions that would be very hard for a normal search to answer. what if i asked him something like, “were you scared?” oh, wow. it’s come up with an answer. william stanley berry states he wasn’t scared after witnessing a plane crash saying, “no, because western dorp said to me, “well, berry, you know the rule.” what’s the rule? after a crash, the personnel involved had to immediately go on another flight. understanding what these people went through helps you to connect better with your past. that’s incredible, isn’t it? that it’s able to find these anecdotes and stories so quickly. and within a few minutes of just asking very simple questions about william stanley berry’s life, we’re down this rabbit hole

1:35 pm

trying to find more and more. if you had to go in and to imperial war museum and manually look for something, it could take you weeks and months. this can take you seconds now. perhaps we can ask him if he’s got any funny stories. yeah, let’s ask him that. the man with the dirty boots was made to empty his pockets, revealing a comical assortment of items. berry laughed and was promptly put in the guardroom himself for laughing on escort. worse crime than the prisoner. recording: it was so funny that i burst out laughing. information was kind of locked up within this audio file, but now it’s transcribed and now it’s searchable using a large language model. all of a sudden, you’re able to not only discover this information, but pull it together and understand it in the context of other things. the whole point is to better educate visitors about history - but, remarkably, the ai has been able to point to new findings that weren’t initially noticed. there are lots of soldiers that served in the first world war

1:36 pm

on continental europe, and they referred to wipers a lot. for ages we were thinking, “what’s this?” “we’re being sent to wipers.” what does what does that mean? it turns out they’re in belgium, and the ai realised the context of these interviews and where they served and figured out they were talking about ypres, the city in belgium. the tool is also challenging history and alerts when it thinks something is questionable. if someone’s uncertain about what they’re saying, it can understand that and flag that. so the reviewer or the researcher looking at that material can go and do their own investigation. and the ai is getting smarter by the day. our ai models are constantly learning from what they’ve already done and refining their learnings, so they’re able to understand accents a lot quicker - and in this particular case, these accents, these regionalised dialects are perhaps not around any more. so actually, we’re just learning directly off the imperial war museum’s data. recording: six of the 18 had to be written off

1:37 pm

because when they landed, they crashed. really? yes. however, like many conversational ais out there, it’s a work in progress. if you ask it the same question twice, it can give you a different response, which makes you wonder, could there be a danger of rewriting history here? it’s really important that when we’re talking about war and conflict in particular, that that information is really accurate. so, going through this process, i think we were aiming for maybe more than 95% accuracy. and what we’ve achieved so far is 99% accuracy, which is consistently better than the human error rate. that 1% that is really, really important, where we need a human expertise from imperial war museum to overlook it and make sure that we are keeping history exactly as it was said in those audio files. the online platform goes live to the public this autumn,

1:38 pm

with almost all of the searchable histories being translated into various european languages using ai. this is another example of how artificial intelligence is increasingly integrating itself into not only our present, but now our past. it’s three hours before the flight. you’ve got your passport in hand and your bag just disappeared on a conveyor belt, hopefully making it to the plane. it’s the experience of millions of us every year, and our thirst for travel is making airports like here in amsterdam look to new technology. airports might seem like obvious places for automation,

1:39 pm

but with thousands of passengers, bits of luggage and other trucks, introducing self-driving vehicles isn’t without its challenges. so we’ve come behind the scenes here to find out what the future has in store for airport logistics. we tried to dissect various processes, or parts of the airport where we can trial with these new forms of technology, and then if that is proven to be safe and successful, then we can scale to different parts of the airport or different processes. there is so much traffic, planes, people. it’s quite a big challenge having it here. indeed, very challenging to add self-driving technology in this system that is already optimised. and as you can see, it’s very busy already. we don’t believe in a turnkey moment that we just go to bed one day and then wake up to an autonomous airport. so we have to trial, we have to explore new technologies, new use cases to gradually grow towards that point.

1:40 pm

the airport’s latest trial is of an autonomous baggage dolly that can help get our luggage from storage in the terminal out to the aircraft. it’s effectively like any other vehicle on this airport. you say where you are now, go to stand 17, go pick up a bag, take it to the baggage hall, go to one of the belts or belt 24 and off-load it there. and the computer inside the vehicle will determine the route it needs to take, and then go and do it. it has to follow vehicles in front of it, stop at junctions, give way to pedestrians and it does that by using lidar. so we have a lidar sensor here. we also use a number of cameras. so these cameras might be just cctv cameras for like evidence gathering for example - but also we use stereo cameras here, and the stereo cameras are used to identify an obstacle. so, is it something that’s fixed? is it something that could be moved, like a bag, that’s in the way? this is one of the often hidden parts of the airport,

1:41 pm

right underneath the main terminal. this is the transfer luggage hall, where baggage that’s come off of one plane and needs to go to another one is stored on. one of the big challenges is how to store them and how to transport them between the planes - and that’s what they’re hoping the self-driving vehicles might help with. for big flights, luggage is often loaded into these containers, called ulds, and carried on a trailer out to the aircraft once several of them are full. but the design of this autonomous vehicle changes that process. so this vehicle looks completely different to what you see on an airport because the ulds are on the vehicle not towed behind it. once that one uld is ready, let’s get that sent out, load it onto the aircraft. then the next one can be getting built in parallel. so, the idea is, is to look at challenging the current way that operations are done in airports. schiphol airport have also been trialling self-driving vehicles for transporting flight crew and passengers requiring additional help.

1:42 pm

but the road to us routinely seeing this tech in operation is long. what is the timeline to actually making the airports autonomous? the timeline for a fully autonomous airport is set on 2050, 5-0, but i’m very surprised in the last two years how far technology has matured. however, integrating that safely in an already very crowded operation might still require us to wait until 2050. with self-driving vehicles becoming more common in other industries and even on our roads, that timeline might feel long. but this is a highly regulated environment, and one of the toughest challenges is convincing authorities this innovation can be brought in safely and ultimately keep airports moving so our flights depart on time.

1:43 pm

1:44 pm

could the next medical breakthrough be found here in dartmoor national park? we’re here to meet a team from basecamp research who thinks it could be. we take the soil sample and put it in a sterile bag. we label it with a unique identification number so that we will know, when we get back where it was taken and which metadata that lewis is collecting belongs to it. this work is part of the team’s ambitious project to build the world’s largest biological database. ph is 3.52. the aim? to unlock new scientific discoveries from nature,

1:45 pm

such as new materials or even medicines. with so much still undiscovered they’re sampling far and wide. what have you found already? ten billion new genes! they laugh from not just dartmoor obviously, but from all the places we’ve been and all the areas that our collaborators have found, have sampled. with so many genes, basecamp has created an ai model to sort through it - but first, the samples are sent to their lab in london. we use extraction protocols to take out all of the dna that’s available in these environmental samples and sequence them. then their ai model can process the data in the dna built specifically to find combinations in the genes, to find new applications. so we do both discovering existing enzymes in nature that have useful functions that can be used in the industry, in agriculture, for pharmaceuticals.

1:46 pm

but we’re also aiming to use ai to generate novel proteins and novel stretches of dna that have functions that don’t exist in nature yet. we’re going to go and extract some dna now. it all started with a trip where founder glen gowers realised how much was undiscovered. so we’ve taken a sample here from vatnajokull, and we’ll be sequencing that back on the ice cap in our base camp with our portable lab. we started the company on an ice cap. we were in iceland on a month long expedition, aiming to become the first team to do fully off grid dna sequencing in the arctic. hey! there we go. we’ve sequenced 100,000 individual strands of dna from a sample near the ice cap in our tent, using solar power. one fact that really shocked us is when we looked at the data, the majority of it was completely new to science. there are currently tens of thousands

1:47 pm

of untreatable diseases with existing tools. we think we can make a dent in that. although smaller, databases like basecamp’s have been open source for researchers to access. basecamp’s data, however, is a commercial endeavour. rather than just opening up our database as raw data, we found that the tools really weren’t there to make use of it in the best way possible. so we build all of our own ai models and with other collaborators, build ai models too, that really are fully optimised for the data sets that we have here. so, you’re in 27 different countries, many of these are in the global south. how do you ensure that you’re not just extracting information and the nature that is these countries? well, historically these activities have been extractive, and that’s really at the core of the problem. to date, there haven’t been economic partnerships between the users of this data and the countries where this data comes from. and that’s really something that, at the core, we wanted to fix from day one. so we build partnerships in all the countries we work with. we share royalties back with every country that we work with

1:48 pm

when revenue is generated from data that comes from their location, and we also engage in short term benefit sharing. so we share data, we build labs, we train people on using these labs, and they’re free to use those labs for anything else that they might like. for field workers around the world, their samples could lead to the next breakthrough. well, one thing that we’re talking about as a company is, for example, contributions to medicines that can cure certain cancers. and that’s not a utopia, that can be done. like there’s been other cancer treatments that have been found in deep sea sponges. and, like, for me, that’s so cool. like, by just sampling more of unknown nature, we might be able to cure some rare diseases or not so rare diseases. the full potential of this research is still unknown, but with so many species, genes and proteins undiscovered basecamp’ research may just be the start of bigger things to come.

1:49 pm

founded in the 1920s as the research arm of the bell telephone company, bell labs became known as the idea factory, a place where scientists, engineers and mathematicians were given the freedom to experiment. the result? some of the greatest inventions of the 20th century. we’ve always solved the hardest problems of communication, networking between people and more and more between people and machines. if you’ve ever made a phone call, watched satellite television, or even used solar power, then you’ve been touched by innovations that came from this place. i was given a very special tour of bell labs museum and archive by lead archivist ed eckert.

1:50 pm

this was the first commercially successful fax service in the united states, and this is 1924, and the quality was so good it made it a successful product. the image here was made from a negative from this, and you can see it’s a high quality image. the technology behind this, the real reason we went for this, is because we wanted to use the phone network for more than just voice. so the next idea is - could we use it to transmit images? and that’s why this was a commercial success. this is from 1964. this is our first attempt at video phone call service. it has an integrated camera, it has a screen so that you could see the person that you were actually calling. there’d be a telephone right next to the set, pick up the phone, be able to see the person. you could, even in later models, share a document. that’s amazing. why did i not grow up with these? well, the service worked very well, but it was extremely expensive. so it was ahead of its time, and a little bit too much money. so if things had been different, we could have been working from home and video calling 40, 50 years ago? that’s right.

1:51 pm

this is the first transistor invented at bell laboratories all the way back in 1947. wow. and without this, so much of modern technology wouldn’t exist as we know it. this revolutionised all electronics. and today, in your cellphone, typically you might have 15 billion transistors. that’s incredible. that’s crazy that something so unassuming can be so transformative. this got us one of our ten nobel prizes. newsreel: as it hurtles through space at a low point of 600 miles... up in space, telstar, a technological marvel and the start of the global broadcasting era. that’s actually an original 1962 telstar 1. telstar 1 was the first satellite that could transmit video across oceans. the other one of those is still floating around in orbit. nine times a day, and it’s considered space junk. but back in 1962, it transmitted the first video across the ocean.

1:52 pm

it carried speeches from the us president, and it carried the first sports event. we actually broadcast for a few minutes, a baseball game. but what’s bell labs cooking up now? from the infinite darkness of space... ..to the wonder of light. this is photonics - the science of using light to move vast amounts of data at frankly ridiculous speeds. bell labs helped invent the field. now they’re pushing it further, experimenting with new innovations aimed at vastly increasing bandwidth and resilience. tell me a little bit about this equipment that’s in front of us here. yeah. so, what this is, is our set-up for generating very, very high speed electronics signals as a switch, an optical switch. so we want to kind of want to turn the light on and off and modulate it very, very fast. so what’s the difference with this kind of set-up here to what i might have with a fibre connection

1:53 pm

at home, for example? so a normal fibre at home to your house might be something like 100 megabits per second, whilst the amount of data that a set-up, that this could send, would be something like four terabits per second. whoa. so, how many times does that light switch on and off, sorry? 440 billion times per second. per second? so, yeah, something like 98% fidelity. one of the wildest concepts under development - lunar cell networks. a vision to bring mobile connectivity to the moon and back, enabling seamless communication for lunar missions and habitats. nasa wants to put humans back on the moon. bell labs wants them to have 5g. in march 2025, as part of intuitive machines’ im-2 mission, nokia bell labs lunar surface communications system landed on the moon aboard the athena lander, powered up and then transmitted operational data back to earth.

1:54 pm

so, tell me a little bit about this here. what is it? so this is what we call a network in a box. it’s basically your entire cellular network in something that looks like a small pizza box. this same unit was recently deployed to prove that we can build a cellular network and deploy and operate this on the lunar surface. so there’s one of these on the moon right now? there’s one of these on the moon right now. there’s a couple of other elements. this is what i would call a lunar smartphone. it doesn’t have a screen, doesn’t have a keyboard because it wasn’t intended for astronauts, like this rover that’s built by lunar outpost, and this is the communication module. can i hold it? yes, of course. oh, wow. it’s quite heavy. that goes into the rover, and it connects to the cameras and the scientific experiments that are on the rover and sends it to the lander. and from there, it’s transmitted back to earth. why do we want a cellphone network on the moon? there are no people there. there are no people yet. but we are going back.

1:55 pm

and ultimately we don’t want to go back like the apollo days where we have two astronauts land and come back a couple of days later. we’re thinking of permanent or semi-permanent presence on the moon, using that as a springboard to go to mars and get the fuel, for example, to go to mars. so when you think about all these missions and that economy, it’s absolutely clear you need advanced communication capabilities. from the tiniest transistor to tomorrow’s moon-ready cell network, bell labs has been quietly shaping the way we communicate for nearly 100 years - and if history is any guide, the next big thing could be happening here, right now.

1:56 pm

1:57 pm

1:58 pm

1:59 pm

2:00 pm

live from london. this is bbc news. two men, both british nationals, have been arrested after a mass stabbing on a train from doncaster to london. at this stage, there is nothing to suggest that this is a terrorist incident. i am alive at huntingdon station where the trainer still is is as investigations continue 11 11 people were injured following following the stabbing spree, to to remain in hospital with life-threatening life-threatening injuries. the bbc speaks to eyewitnesses, who describe the terrifying moments on board the train. i look at my hand and it is covered in blood and i look

0 Views

info Stream Only

Uploaded by TV Archive on November 2, 2025