now, sport and for a full round—up from the bbc sport centre. hello from the bbc sport centre. we start with another busy night of european football and in particular the europa league where manchester united, viktoria plzen, rasmus hojlund came off the bench to score twice as the visitors took all three points, after back—to—back defeats domestically. the result lifts ruben amorim's side into the top eight positions that would automatically see them qualify for the last 16. hojlund has now scored five times in europe this season and united remain unbeaten in the competition. busy night in the europa league. these are the that we are taking a close eye on. the

second family but am is an all british affair between rangers and telling him. it is all this as things stand. ten points from their five as things stand. ten points from theirfive games. as things stand. ten points from their five games. you as things stand. ten points from theirfive games. you can see with some results with united winning and a big win for roma. it was to have into against malmo and in the fourth game one point ahead. the former barcelona youngster marc guiu was the star of the show as he helped chelsea into the knockout stage of the uefa conference league with victory against astana in kazakhstan. a much changed squad of young and fringe players made the 16 hour round trip to almaty for the match, which was played in temperatures of around minus 11 celsius. and chelsea's ability to field an almost entirely different starting line up to the team that beat tottenham was on show as they got the job done

with all the goals coming in the first half. lots more games going on in the conference league tonight — far too many for us to show you — but a few games of note. fiorentina put seven past lask whilst elsewhere hearts suffered a 2—0 defeat away to copenhagen. northern ireland side larne were beaten for the fifth time in the competition, 2—0 by dinamo minsk. shamrock rovers have the lead whilst the new saints are losing. keep up to date with a full list of scores on the bbc sport website. to the women's champions league where already—qualified manchester city are looking to secure top spot in the table they're playing st polten and it's currently goalless while barcelona are leading hammarbyi—o. while in the early games, two goals from alessio russo helped arsenal to a 3—1 win at valerenga.

while bayern munich beatjuventus 3—0. ——a—0. after their defeat to juventus last night, manchester city's attention turns to playing manchester united on sunday. for city, it's seven defeats in ten, just one win, the worst run of manager pep guardiola's managerical career but he believes the side is still playing well despite the results. i thought we played really good. i am asking what did we do today? we played really good. i am living the good and bad moments. iam good. i am living the good and bad moments. i am trying to find a way and do it but the game was there so that i am incredibly honest about knowing when we play good and not. some

uolf to when we play good and not. some golf to bring _ when we play good and not. some golf to bring you — when we play good and not. some golf to bring you before _ when we play good and not. some golf to bring you before i - when we play good and not. some golf to bring you before i go. - and andy sullivan hit an eight under par round of 64 to go a shot clear of fellow englishman. alex fitzpatrick and home player casey jarvis after the opening round of the alfred dunhill championship in south africa. two of sullivan's four previous european tour wins have come in the country. and that's all the sport for now. thank you. you are watching the context. it is time for al decoded. welcome to the programme. if you were watching us earlier in the summer, you may recall we brought you a special programme from wimbledon, where the all england club has has been adapting al, to improve its line calls. next year for the first time in the championships history, there will be no linejudges. 18 match courts will be installed with automated electronic line calling. so what about football? how long before we do away with the assista nt referees ?

new ai powered semi—automated technology is being trialled by the premier league this season. it is thought the new technology will reduce the time taken for offside decisions by up to 31 seconds. it will also provide a clearer image for us the spectators. technology in football, var, is a controversial issue. but is it the technology, or is it the people using it. some clubs in the premier league are already using artificial intelligence on tactics, performance, to evaluate injuries, and to improve supporter engagement. joining us tonight from berlin our regular contributor and ai expert stephanie hare and here in the studio the bbc�*s tamzin kraftman who's been looking at this new ai technology and how it could impact the role of referees in the premier league. just so we are very clear about what this will be about, in terms of the technology, is semiautomated of side technology different to what we

currently have with the ar? yes. if you are looking at ver, video assistant referees, that is where a team of officials are looking at video, so they have independent access to footage and multiple screens and can go back and look at a call if there is a case where the on field ref isn't quite sure needs a second opinion. we are talking about the next type of technology part of the show which is involving cameras that are coming underneath the roof of the football stadium and linked with sensors, there is a sensorin linked with sensors, there is a sensor in the ball. everything is there to create a play for play, a point virtual digital twin of everyone in the pitch. so they are able to see with greater and faster precision or as var sometimes takes ages and frustrated fans, this new technology is faster and should

keep the flow of the game going with critical decisions. you have been _ with critical decisions. you have been looking - with critical decisions. you have been looking at - with critical decisions. you have been looking at it. i with critical decisions. you have been looking at it. how file do you think the premier league will go with it? it has been looked _ league will go with it? it has been looked at _ league will go with it? it has been looked at in _ league will go with it? it has been looked at in the - league will go with it? it has been looked at in the last. league will go with it? it has| been looked at in the last few years— been looked at in the last few years for— been looked at in the last few years for example in the world cup in — years for example in the world cup in 2022 and euros last yeah _ cup in 2022 and euros last yeah in _ cup in 2022 and euros last year. in the premier league, we will see — year. in the premier league, we will see the _ year. in the premier league, we will see the most powerful system _ will see the most powerful system yet or so they say. but how — system yet or so they say. but how will — system yet or so they say. but how will it — system yet or so they say. but how will it impact to the assistant referees that they are seen on the pitch? we do not think— are seen on the pitch? we do not think that similar to wimbledon that lines will be abolished but i think the biggest question we have to ask is how— biggest question we have to ask is how it — biggest question we have to ask is how it will impact their jobs _ is how it will impact their “obs. ., �* , .,, is how it will impact their “obs. ., �*, ., jobs. the world's most famous sort is jobs. the world's most famous sport is about _ jobs. the world's most famous sport is about to _ jobs. the world's most famous sport is about to get _ jobs. the world's most famous sport is about to get a - jobs. the world's most famous sport is about to get a face - sport is about to get a face lift in the form of semi—automated

offside technology. the ai powered system is set to come to the premier league this season, after being featured at the world cup in 2022 and the euros just last year. the winners of the contract to develop the technology for the premier league, genius sports, are keeping their new system under wraps. but to understand how semi—automated offside technology works, i visited ai data platform encord. they place cameras all around the stadium. it's been used already in the euros and the world cup, so around about 12 cameras and essentially using those cameras, they're able to reconstruct the positions of the players and of all the different points on the players. so reconstructing, for example, shoulder positions, hand positions, etc, all in sd and then processing that with al. so can we actually see this technology in action? yes. here what we have is a video of a football match. and you can see that we have these masks around the players. they are essentially tracking them, identifying them, giving them kind of unique ids so the system can understand who's in what team, where they're positioned,

and critically, as well, we have the ball being tracked here in this camera. now, obviously, this is a single—camera view. and when the models are working you know, on on the pitch, they're processing these they're processing these 12 different cameras. and having this view from every 12 different cameras. single angle allows them to reconstruct this 3d to reconstruct this 3d representation of the game. representation of the game. what will they have to do, what will they have to do, then, if semi—automated offside then, if semi—automated offside technology doesn't actually work? technology doesn't actually work? so, it has failed in the past. so, it has failed in the past. so for example, so for example, for the automated goal for the automated goal technology, that's been used technology, that's been used on 9,000 games with one on 9,000 games with one error, but at that point error, but at that point it's up to the referee it's up to the referee to actually make that call. of the most renowned to actually make that call. i'm here at the home i'm here at the home of football, where many of football, where many of those cameras are set of those cameras are set to be installed for to be installed for the fa cup final in may. the fa cup final in may. if the fa manage to roll out if the fa manage to roll out the technology in time, the technology in time, semi—automated offside semi—automated offside technology is hoped technology is hoped to make the role of to make the role of the referee easier, allowing the referee easier, allowing them to make quicker them to make quicker and more accurate decisions. and more accurate decisions. but what do referees think? but what do referees think? earlier, i spoke to one earlier, i spoke to one

of the most renowned

guidance and mentoring to referees of all levels. they say the changes brought on by semi—automated offside technology are welcomed. i don't think any referee has a particular problem with this. the same with tennis, as it's going to be very good atjudging whether something is over that line or drawing a line, or deciding where that line should be, and it's just going to speed up the check of that process. and technology is very good at doing that. that's because it's a very simple black—and—white task, and in those sorts of circumstances, technology is there to augment, to help. we've still got the umpire in the chair. and for us, it's always going to be assistant referees, referees, fourth officials and the guys in the var booth, as well, looking at things. thejobs of referees in the premier league appear to remain secure for now. but as the sport continues

to experiment with what ai can offer, the focus could soon turn to what other roles technology could replace. i would think supporters would welcome this with open arms because having been to many primo games in my lifetime, one of the biggest frustrations with var is you have to wait in order to celebrate. 31 seconds is an age when you just scored a goal and one to celebrate. so if there can be made instinctively and communicated to the fans in real time, that must be a bonus —— can be made instantaneously. i5 must be a bonus -- can be made instantaneously.— instantaneously. is hoped to be makin: a instantaneously. is hoped to be making a huge _ instantaneously. is hoped to be making a huge difference. - instantaneously. is hoped to be | making a huge difference. there is a lot of talk about quicker response time but a lot of the decisions are made with the var. so there is a lot that we have to wait to see that the premise here is that it will be a lot more accurate and essentially, reduce a lot of

the upset that was happening but both fans and players in regards to the long waiting times. ., , times. the other thing is if ou do times. the other thing is if you do not _ times. the other thing is if you do not have _ times. the other thing is if you do not have to - times. the other thing is if you do not have to have i times. the other thing is if you do not have to have a | you do not have to have a linesman running down the line then presumably, you can put these assistant referees in different positions on the pitch which might also help the game. the issue about mine technology at wimbledon is that it is very black—and—white. it is either out or in. the slimmest duets offside, —— the same is true with our side. can you see a scenario where ai could be developed to make subjective decisions which i think a lot of fans have an unconscious bias within them? can you see a scenario where it can make a decision on a free kick or penalty? i can make a decision on a free kick or penalty?— kick or penalty? i think it would be _ kick or penalty? i think it would be more _ kick or penalty? i think it would be more fun - kick or penalty? i think it would be more fun to - kick or penalty? i think it | would be more fun to ask kick or penalty? i think it - would be more fun to ask what would — would be more fun to ask what would al — would be more fun to ask what would ai do would be more fun to ask what would aldo with those historic

moments _ would aldo with those historic moments in football that have been — moments in football that have been so — moments in football that have been so controversial like the hand — been so controversial like the hand of— been so controversial like the hand of god moment with maradona, david beckham being sent off— maradona, david beckham being sent off in the world cup. well— sent off in the world cup. we'll have to train a on past data — we'll have to train a on past data -- _ we'll have to train a on past data -- we _ we'll have to train a on past data —— we will have to train ai data —— we will have to train aland — data —— we will have to train aland pass data. how would ai and pass data. how would they— aland pass data. how would theyjudge past because and how will it_ theyjudge past because and how will it affect things going for it? i— will it affect things going for it? ithink will it affect things going for it? i think i will it affect things going for it? ithink i will be will it affect things going for it? i think i will be fun seeing— it? i think i will be fun seeing how ai will put a human referee — seeing how ai will put a human referee to _ seeing how ai will put a human referee to raise their game. yeah — referee to raise their game. yeah you _ referee to raise their game. yeah. you seemed a bit sceptical, if you don't mind me saying, in terms of the fa cup final. but i think of stadiums in the premier leagues that have done that is there a problem with rolling it out and make the technology for it to the stadium we have? if remember correctly if some stadiums don't have var

technology they want have it for a whole tournament. i believe it happened with a championship final a few years ago. there is a very good question and i think it is 12 cameras need to be installed and if you think about the amount of stadiums, it needs to be installed in so many stadiums, may be the targets of the final could be out of reach. i the final could be out of reach. ., ., the final could be out of reach. ., , . . the final could be out of reach. ., ., ., reach. i would pay an extra pan each week _ reach. i would pay an extra pan each week for _ reach. i would pay an extra pan each week for a _ reach. i would pay an extra pan each week for a proper - reach. i would pay an extra pan each week for a proper offside | each week for a proper offside decision —— an extra pound. coming up... we will be speaking to brian hall, founder and head of artificial intelligence at alphaplay ai, a powerful ai platform used by the global sports industry. he is standing by to tell us how far he think this technology can go. around the world and across the uk, this is bbc news.

if you have used artificial intelligence in your work, then you will know already that it offers two distinct benefits. firstly, it analyses a significantly larger pool of data, in real time. second, it delivers insights

in clear, actionable language. if used correctly, it can improve decision—making. so you can see clearly how it would be adapted to sport. rather than replacing the experts, the coaches that clubs have retained, you could probably view it as a more symbiotic relationship between technology and the professionals in sports. the expert in this field is brian hall, founder and head of ai at alpha play which orginally was conceived for baseball. it has proven even more effective in the complex team sports like hockey, football, and the nfl. let's talk to stephanie because we have talked about how you can train referees. in terms of transfers, rehabilitation, the sky is the limit when it comes to ai. ., ., 4' to ai. you would think the potential— to ai. you would think the potential for— to ai. you would think the potential for coaching - to ai. you would think the potential for coaching are | potential for coaching are preventing are coming back from injury, all of the data we have an our bodies, biometric data like heart rate, blood tests,

to check and see how players are doing, it can be modelled over time to know exactly when to optimise for each player, when to bring us and put them on the pitch, who does good and bad under pressure. then you have to coach and train them. so if you look at a football player today and compare them back to the 19705, they are much fitter and have much better diet which is obvious. they recover from injuries faster, as well. imagine where we will be in 20 to 30 years' time with personalised medicine and coaching. i5 time with personalised medicine and coaching-— and coaching. is that where machine — and coaching. is that where machine learning _ and coaching. is that where machine learning comes . and coaching. is that where | machine learning comes in? and coaching. is that where - machine learning comes in? as you say, you feed it with historic data and games, play hi5toric data and games, play formations, how player5 react under pressure and in certain games. under pressure and in certain game5. have you under pressure and in certain games. have you seen under pressure and in certain game5. have you seen evidence in other sports where that does make an enormous difference? well, you can look in sports and in the military or even in

anything where you are doing scenario training. artificial intelligence would be fabulous intelligence would be fabulous in a coaching environment, and football match live, for instance, you could have the whole huddle of teams running different scenarios of a they want to do in the next two minutes and see which one the ai says is the most likely to win a goal or equalise or defend against the other team if you are playing for time. it is like having a coach that can run so many different scenarios fa5ter run so many different scenarios faster than any human but it will still lack things like intuition and personal knowledge. machines cannot read everything. you're notjust a heart rate or how many calories you burn are callers —— or because you score. you still need to know the player. there will be coaches _ need to know the player. there will be coaches out _ need to know the player. there will be coaches out there - need to know the player. there will be coaches out there who l will be coaches out there who are sniffy about using ai in day—to—day operations around the team because they may feel

that it stifles the experience that it stifles the experience that they bring or the intuition to change tactics. actually beginning to the referee that you spoke to in your 1