When Adam Haar Horowitz took to the stage at a conference dressed as a lotus flower, he raised eyebrows.
Then, when he started hitting computers and making gong noises, jaws dropped.
He was acting out a dream he had recently had to illustrate how our night-time fantasies can influence our waking lives, and how technology can help us access them.
It is a subject close to Mr Horowitz’s heart.
“Dreams are such a strange, murky, inaccessible space and there is so much poetry, metaphor and analogy in them,” he told the BBC when it visited him at the Media Lab in the Massachusetts Institute of Technology.
“The idea that you can take something concrete – a technology – that can help you access that poetic and metaphorical side of your own cognition is really exciting.”
To achieve this he has invented a hand-worn device he calls Dormio.
It collects biosignals that in turn track transitions in sleep stages – such as a loss of muscle tone, heart rate changes, and alterations in skin conductance.
The goal is to study a particular stage of sleep – the period between wakefulness and deep sleep, known as hypnagogia.
It is a period of slumber which has fascinated scientists and artists for hundreds of years.
Thomas Edison, Nikola Tesla, Edgar Allan Poe, and Salvador Dali all attempted to access this state by napping with a steel ball or similar object in their hands. When they fell into deeper REM (Rapid Eye Movement) sleep, they would drop the object, waking themselves before they forgot the hypnagogic imaginings.
The Dormio gadget is connected to a smartphone app or robot, which speaks word prompts to the subject as they slip into deeper sleep. These words can be used to influence their dreams or to knock them back into lighter sleep.
“We have found that in the subjects we tested, those words reliably entered the hypnagogic dreams as dream content,” said Mr Horowitz.
“After this slight wake-up, we initiate a conversation about dream content with users via the Jibo social robot and record anything that is said, as hypnagogic amnesia is reported and we don’t want people forgetting their useful ideas.”
After this conversation, the system lets users drift back towards sleep, interrupting again when the biosignals suggest they are falling into deep sleep.
“This is done repetitively to intercept dreams and extract dream reports,” said Mr Horowitz.
So far it has only been tested on 15 people but the hope now is to widen the trial and eventually make it commercially available.
“I see a future in which sleep is more useful and more accessible to us, where we understand it better and where we have more possibilities for influencing it and extracting the cognition that goes on inside it,” Mr Horowitz told the BBC.
Learning what goes on in dreams will help the wearer consolidate their memories, improve their learning as well as influencing how they dream, he added.
“There is this overnight period which we largely ignore but which is so crucial to how we remember, how our emotions balance out during the day, how we deal with trauma and how we learn – and we just kind of ignore it; that is very weird to me.”
We spend a third of our lives sleeping but precious little is really known for sure about what happens when our heads hit the pillow.
Sigmund Freud famously believed that dreams allow us to fulfil our greatest desires. Other theories suggest that dreaming is a way to process and understand emotions and events, perhaps a method of storing less important memories or even a way of training our brains for threats that might await us when we are awake.
It is generally agreed that people dream between four and six times a night and forget about 90% of the dreams they have had.
Antonio Zadra, a psychology professor at the University of Montreal, believes that dreams are designed to be forgotten.
“We spend six years of our lives dreaming, of which we only remember a small sliver. If we were meant to be remembering more, that would be a colossal waste of time,” he said.
While the dreams in deep REM sleep are, according to Prof Zadra, “impervious to external attempts to influence them”, hypnagogic dreaming is easier to influence, although he is not convinced that this needs the help of new technology.
“The metal ball works really well too. The MIT system is just fancy-made tech that is doing the same thing and could be over-complicating the process.
“But, if someone is interested in exploring the images they have when they are asleep, then a device like this could be useful.”
Mr Horowitz admits there are ethical questions that need to be thought about before engaging with this enigmatic part of our daily routine.
“How much of yourself do you want to access?” he ponders.
“How much power do you want to have there? How much do you want to be careful about messing with your own biosignals?”
As well as creating the tools that allow people to understand the different stages of sleep, he plans to share some of the vast trove of literature on the subject “in an accessible way” so people can decide for themselves what they use their dream-interrupting technology to achieve.
Ultimately, he sees it as a part of the journey of learning about oneself.
“The waking you will be more connected to the sleeping you and that connection between the two is a form of self-discovery,” he said.
I leap from the battle bus and land in the middle of what appears to be a deserted island, carrying only a pickaxe as a weapon. As I start running, I hear the sound of gunshots in the distance. Almost immediately I’m dodging bullets and suddenly I’m dead – shot by one of the other 99 players on the island.
I’m playing Fortnite, the video game phenomenon. According to developer Epic Games, more than 125 million other players are also in its grip.
It’s the first time I’ve properly dedicated myself to a game since Tomb Raider in the 1990s, when I was a 20-something rather than the 40-something I am today.
Battle Royale is Fortnite’s most popular mode. One hundred of us collect weapons, build defences and track each other down until there is only one person left alive. And every few minutes, the map shrinks.
I launched straight into my first game with little preparation, on a smartphone.
At first, I found it hard to work out what I was supposed to be doing and my not-so-great eyesight struggled with the small screen.
After a few minutes, I gave up and downloaded the game on my iPad and later on my PC.
Although the bigger displays helped, I did not last long.
During my first few attempts, players would appear out of nowhere and kill me in an instant – leaving me frustrated that I hadn’t even noticed them approaching.
After a few games on the tablet, I worked out that I had to hunt for better weapons – and I needed to collect materials to build forts to protect myself.
It wasn’t until I played on a console belonging to a friend’s teen son that I really felt like I got into my stride.
Using the Xbox controller, I could more easily pick up the weapons I found on the map, which I had often missed while playing on the tablet.
On the large TV screen, I could also get a better sense of where the other players were – and what they were doing.
And I was given some expert coaching from 14-year-old Archie, the son of another friend.
“It’s easier for me to navigate having a [console] controller than using a PC mouse or keyboard,” he told me.
Although he added that Ninja – Fortnite’s top Twitch streamer – played on PC.
Even as I collected better weapons, I was lucky if I managed to survive beyond the final 60 before being eliminated by a more experienced player with better skills and weapons.
But it was easy to understand Fortnite’s appeal.
I felt compelled to try to improve on my position and survive for longer.
I found myself picking up the tablet while watching TV, to launch into a quick game.
Before I knew it, I had spent an hour playing.
Once you’re eliminated, you can choose to watch the person who beat you, to see how they progress in the rest of the game.
Many of the players have spent money on in-app purchases such as “skins” that change the appearance of a character and “emotes” that let characters perform ridiculous dance routines.
When I was killed by a jester who then danced a jig, I was less than impressed.
I’m not ashamed to say I’m a hider. My player strategy consisted of trying to avoid other players.
If I saw someone nearby, I would run and hide in one of the many bushes or buildings.
My sophisticated strategy is unlikely to lead me to victory – I have never survived beyond the final 17 players.
Archie said he had won 10 games out of about 200 and played against his school friends online. “You get a real buzz with a Victory Royale,” he told me.
I certainly don’t think I’m ever going to be good enough to win. But I keep dipping into the game to try to beat my record of 17th place.
Each time I’m eliminated by some stealthy player who catches me unaware and shoots me in the back, I find myself shouting in frustration.
I just do not see them coming and each time my reaction is the same – a groan.
I’ve already lost count of how many games I’ve played so far.
Some parents think their children are addicted to the game and would happily spend all their time playing it.
Ruth Kieran told me her two sons, aged 11 and nine, played online when they got home from school.
“They would spend 24/7 playing Fortnite if they were allowed to,” she said.
“We try and limit our boys to just one hour in the evening after school and a couple of hours at the weekend.
“But it is really hard when they are playing against friends [whose] parents are not as restrictive.”
I’m not sure Mrs Kieran would be too impressed with how quickly I’ve managed to succumb to the game – and I’m not even playing against anybody I know.
I’m literally just trying to beat my last position.
Some experts say parents do need to take more responsibility.
“There are some parents out there who use screen time as a babysitting aid,” said Prof Mark Griffiths, a psychologist and director of the International Gaming Research Unit.
However, he added that computer games were not necessarily bad news.
“When parents contact me complaining about their children spending too much time playing computer games, I ask them, ‘Has gaming interfered with their education or chores or are they not being physically active?’
“When they answer, ‘No,’ I tell them that it is normal and they need to stop worrying.”
Archie suggested playing Fortnite was “like a virtual play date”.
“Parents don’t need to worry where their kids are, because they are at home playing online with their friends,” he said.
As my two-week experiment of playing Fortnite was about to end, I spoke to my 10-year-old nephew.
He had downloaded the game to his tablet the night before, to his mother’s dismay.
She said he had come home from school pleading to be allowed to play Fortnite, because all of his friends were playing it too, despite it not being recommend for under-12s.
I definitely scored a few “cool auntie” points when I told him about my experiences on the island.
But when my “sensible adult” mode kicked in and I warned him not to spend his money on in-app purchases to customise his avatar, he said all his friends had skins.
I don’t think I’ll be getting into any Fortnite battles with my nephew or any of the other school-age children of any of my friends, who found the idea of me playing Fortnite highly amusing.
But to be honest, it is likely I will go home this evening and play another game or two.
By Chris Foxx, technology reporter
An AI-powered search tool is being used to monitor fake news and polling station problems in Mexico’s elections.
Krzana is being used by the Verificado 2018 initiative which was set up by Mexican media to challenge fake news.
The run-up to Mexico’s election has been deadly with more than 130 people killed since campaigning began.
Fearing more violence, Verificado plans to use the tool to react quickly to propaganda on social media and violence aimed at voters.
Verificado has been set up by more than 90 separate organisations including publishers, media groups, NGOs and universities and has enrolled social media giants Facebook and Google to help monitor the election.
The sharply contested election has been marked by widespread use of bots and trolls to spread fake stories on social media.
Krzana will be used during the weekend of voting, said Toby Abel, Krzana co-founder, to help media groups pick out and react quickly to bogus information.
The tool has been taught to scan and analyse text, video and stills on social media to spot items and messages that relate to candidates and the election.
It can do this faster and more comprehensively than a human could, said Mr Abel, but the machine will not be left to decide on whether what it captures is fake or not.
“Working out if something is false or not is something that a good journalist can do very quickly,” said Mr Abel. “It’s unbelievably difficult to teach a computer to do the same thing.”
Once the bogus stories have been spotted, media groups plan to write messages that debunk claims or show who has put them together.
Reaction speed was vital when countering fake news, said Quin Murray, Krzana’s other founder.
“It’s about getting in there before the original has had chance to spread,” he said, adding that the key was tapping into the same lines of communication that help speed stories people want to read across social networks.
“That speed is so important because if someone reads a fake story and logs off that could be the closing thought that they are left with,” he said.
Krzana will also look for information about polling stations on the day of voting as many could become targets for actors looking to influence the election by dissuading voters from turning up.
“There’s quite a large fear that there’s going to be fake information thrown around about attacks and gunfire at polling stations where one party thinks they are going to get more votes,” said Mr Murray.
Media groups had people ready to travel to stations to verify claims of violence or report if they were trouble-free.
Mexico’s general election takes place on 1 July and will see the country elect a new government and president. Long-time activist Andres Manuel Lopez Obrador is the front runner and is expected to win the presidential contest.
Owners of Nissan’s new electric Leaf say they were given misleading information about the car before buying it.
They say charging the Leaf can take three times longer than claimed on Nissan’s website.
Others are unhappy that the range on a single charge is not as good as the 235 miles (378km) they were promised.
Nissan admitted that charging times can vary, but denied there was a problem or that any customers were misled.
The Advertising Standards Authority is now considering whether to launch an investigation into the issue.
As many as 2,600 new Leafs have been sold in the UK, and it was named Electric Car of the Year for 2018 by What Car? magazine.
But drivers attempting longer journeys in the Leaf have found themselves spending up to two and a half hours at motorway service stations to recharge.
Last year, Nissan told prospective buyers that using so-called rapid chargers should only take 40 minutes “in moderate driving conditions” for an 80% charge. They subsequently changed that to between 40 and 60 minutes.
There appears to be no problem with the first two charges on any given day – one at home, and then the first rapid charge en route.
It is only when drivers come to charge for the third time – or the second rapid charge – that some have said they face long waits. Potentially, that could affect any journey of more than 250 miles.
John Weatherley, a company director from the Forest of Dean, loves his Nissan Leaf.
But when he made a 300 mile journey to the Lake District, he found himself waiting for a total of two-and-a-half hours when he stopped to charge for a second time.
“If Nissan at the start had said what the car is capable of, without exaggerating the fact on their website, I’d have been fine with it,” he told the BBC.
“They said they could charge in 40 to 60 minutes, so I believed them. But it’s not true. The advertising is totally misleading.”
When Mr Weatherley wrote to Nissan to complain, he was told that rapid charging was only intended for use once in a journey – something many buyers may be unaware of.
Nissan also told the BBC that charging can take longer than advertised, depending on conditions.
“External ambient temperature, the type of driving you’ve been doing beforehand, and the heat you put into the battery if you’ve been doing successive charges can impact the timing,” said Gareth Dunsmore, director of electric vehicles for Nissan Europe.
He said the battery automatically slows a charge, to preserve its longevity, and to act as a safety mechanism when it gets too hot.
“We make this clear in the owner’s manual,” said Mr Dunsmore.
In some instances it can also be the charger itself that is to blame, he said.
Tony Pitcairn, from Ilkley in West Yorkshire, had problems on a 290 mile drive to Somerset.
He and his wife spent 90 minutes at a motorway services in Gloucestershire.
But Mr Pitcairn was also disappointed by the range of the new Leaf, which he bought specifically for long journeys.
His marketing brochure claimed the car could do 235 miles on a single charge.
But having bought the car, he found the range was actually 155 miles.
“That was a disappointment to start with,” he said.
“So we have, in my mind, been misled twice, because the claimed range on a full charge is not 235 miles. Secondly, nowhere does it say that you will only be able to rapid charge in 40 minutes only once.”
When journalists from What Car? tested the new Leaf, they found a “real world” range of just 108 miles.
Nissan said the original claim of 235 miles was correct under an official means of measurement known as the New European Driving Cycle (NEDC).
However, as carmakers have moved to a different measure – known as the Worldwide harmonised Light vehicle Test Procedure (WLTP) – the range is now officially 168 miles.
Mr Dunsmore advised any upset customers to get in contact: “Come and speak to us if there’s anything you’re not happy with.”
Meanwhile, a number of customers have cancelled their orders.
Matt Beard, from Aberdeenshire, did so after taking a test drive. Eventually he bought the older model Leaf instead, which has a less powerful battery, but fewer problems when charging. In a tweet, he said he was shocked at how bad the problem was.
Others are unhappy with the response from Nissan.
Jonathan Porterfield, of eco-cars.net, who regularly drives electric cars from Leicester to Orkney, was the first to report the issue.
“I don’t want this episode to knock Nissan, but at the same time they need to sit up and take notice,” he said.
“Just telling people you’ve got to wait longer at a rapid charger – it’s not good enough.”
Tesco is testing whether supermarket customers can buy products without visiting a till.
The UK’s biggest retailer has given some staff a smartphone app to use at a store at the company’s headquarters.
They use the app to scan barcodes on the products they want and then pay, with no trip to a till needed.
Co-op is experimenting with similar technology and Amazon already has a checkout-free grocery store in Seattle that is open to the public.
Tesco says its experiment is at a very early stage.
It is concerned that checkout-free stores might be a target for shoplifters and, as it operates on very fine profit margins, that could make the technology unviable.
“If the margin in the business is 2 or 3%, you don’t have to lose much to make it unprofitable,” said chief executive Dave Lewis.
Given Tesco’s size, if it did adopt the technology it would be significant advance for checkout-free shopping.
Co-op is also experimenting with a smartphone app that would eliminate the need to visit a till.
It is running a trial at a store in Manchester and will expand those tests over the summer.
However, retail experts are not convinced that grocers can make big savings from the technology.
“Recent stories about the level of theft at self-service checkouts – something that we’ve had for around 10 years in retail – leaves me to wonder about the obvious savings in checkout-free stores versus a hidden cost in terms of theft,” said independent retail analyst Steve Dresser.
“However, you always follow the customer as a retailer. It has always been clear that customers want as little hassle as possible.”
Amazon’s store in Seattle has a more sophisticated system. Shoppers enter by swiping their smartphone that has the Amazon Go app.
The store uses hundreds of ceiling-mounted cameras and electronic sensors to identify each customer and track the items they select.
Customers can to put any item straight into their shopping bags and purchases are billed to their credit cards when they leave the store.
The Tesco store at its headquarters in Welwyn Garden City is also cashless, which Tesco says has cut the average time spent at a checkout to 45 seconds.
It takes shoppers at similar-sized stores that accept cash an average of 90 seconds at the checkout.
Digital rights campaigners are starting a legal challenge to a US law that seeks to fight online sex trafficking.
The Electronic Frontier Foundation (EFF) says the “poorly written” law can hinder attempts to help victims and prosecute traffickers.
The Fight Online Sex Trafficking Act (Fosta) also trespasses on free speech laws, claims the lawsuit.
The EFF wants the law declared unconstitutional to stop it being enforced.
In a blog announcing its legal action, the EFF said the law needed to be halted because, in its current form, it was harming many people working on behalf of sex workers and victims.
In particular, said the EFF, the vague language in Fosta puts those who call for decriminalisation of sex work, or who try to establish greater recognition for prostitutes and others in the trade, at the risk of prosecution.
In addition, it said, the law undermines established protections enjoyed by websites that host content posted by their users.
Fosta “vastly magnifies” the risk these net firms bear if they choose to run ads on forums dedicated to these sexual professions, it said.
Already net firms including Craigslist, Reddit and others have shut down forums and chat rooms dedicated to the buying and selling of sex for fear of prosecution.
The law has also limited the work of organisations trying to help people who offer sexual services, said the EFF. One such was VerifyHim, which logged descriptions of abusive clients to help workers avoid them.
The EFF’s legal challenge is also being aided by the Internet Archive, Human Rights Watch and Woodhull Freedom Foundation. In addition, two individuals are backing it – one a spokesperson for sex workers and another a masseur who now finds it hard to advertise his non-sexual service.
The US Congress was repeatedly warned that Fosta would not achieve its aims and was likely to foster widespread censorship, said the EFF.
An experimental robot with an animated cartoon face has been sent to the International Space Station (ISS) on board a SpaceX Falcon 9 rocket.
Dubbed Cimon (Crew Interactive Mobile Companion), the device is intended as an “an AI-based assistant for astronauts”.
Cimon weighs 5kg but in zero gravity it will float move around thanks to 14 internal fans.
It is an attempt to find out whether robots and astronauts can collaborate.
To this end, Cimon is equipped with microphones and cameras that help it recognise Alexander Gerst, the German astronaut with whom it will work.
An “offline” button has been designed for Cimon, which allows Gerst to avoid having audio streamed to servers on Earth during moments of privacy.
The rocket whisking the bot on its way to Earth orbit took off from Nasa’s Kennedy Space Center in Florida at 05:42 Eastern time (10:42 BST).
It was part of a cargo shipment weighing nearly 2.7 tonnes that includes food and water.
The cargo is contained in a Dragon capsule that will detach from the Falcon 9 and dock with the ISS on Monday.
“Right now our main mission is to support the astronauts with their daily tasks to save time, because time is the most valuable and most expensive thing on the ISS,” IBM engineer Matthias Biniok told Reuters.
Cimon, which speaks English, was developed by Airbus and IBM for the German national space agency, DLR.
It will be able to give astronauts instructions to help them carry out scientific experiments and it will be able to respond to verbal questions.
“The use of artificial intelligence in assisting space flight is an obviously good use of the technology,” said AI expert Noel Sharkey at the University of Sheffield.
But he pointed out that Cimon would have to prove its worth.
“It is only in the experimental stages at present and if it is not useful, no space agency will want to waste incredible valuable space with it.”
On 13 December, Cimon is scheduled to return to Earth.
Hundreds of fake accounts making “phantom bookings” have recently plagued Singapore-based firm Ryde Technologies, the company has said.
In recent weeks, nearly 300 such accounts have made 2,000 bogus bookings, costing drivers $50,000 (£37,900).
The start-up investigated the problem after drivers complained.
Uber, which offers a similar service, has in the past reported cases of fake bookings in other countries.
Ryde has filed a report with local police who are now investigating the matter, according to Reuters.
“Such manipulations of the app have distressed affected drivers, as their means of livelihood have been impacted,” the firm said in a statement.
“On top of causing drivers to lose money on fuel, these acts severely limit the drivers’ availability to take on more jobs as it sends them driving around in vain when they could have been picking up legitimate riders.”
A spokeswoman confirmed to the BBC that the $50,000 lost by drivers refers to income they would have earned from bookings had they been real.
Ryde added that it has begun conducting user audits to try to eliminate fake accounts and it has encouraged drivers to report phantom booking incidents.
The practice of lodging fake bookings is especially problematic for a young company in such a competitive market, said analyst Hanish Bhatia at Counterpoint.
“It definitely hurts the business,” he told the BBC.
“It creates a very negative sentiment among drivers who have just joined.”
Notably, the Asian ride-hailing market, in which giants like Didi and Grab also operate, is highly “price-sensitive”, said Mr Bhatia.
Firms competing on price may be particularly affected by lost business, he added.
Mr Bhatia also suggested that ride-hailing firms, as they grow, gather data on all drivers and passengers that may reveal when unusual activity is occurring or when fake bookings are being made.
Karamo Brown, a host on the hugely successful makeover show Queer Eye, has vowed to pressure streaming service Netflix into improving its subtitles for deaf and hard-of-hearing watchers after a social media debate over their quality.
Fans of the show took to Twitter to complain about the service misrepresenting, censoring and simplifying dialogue from a variety of shows.
Mr Brown, the Queer Eye cast member who focuses on culture, said reading fan comments had broken his heart.
Reading everyone’s comments breaks my heart. I don’t know how much power I have but know, the next time I’m at Netflix I’m going to bring up this issue internally & wont stop until something changes. Deaf & HOH people should have the same experience as everyone else! #TypoFixedhttps://t.co/AQ4emvgUBv
— Karamo Brown (@KaramoBrown) June 28, 2018
End of Twitter post by @KaramoBrown
Tweets by Rogan Shannon, a deaf Netflix fan, in which he demanded that the service explain why it was not captioning word for word, have been shared thousands of times in recent days.
His tweets claim the subtitles censor profanity and edit dialogue for brevity.
Others accused the service of failing to caption foreign language inserts and correcting distinct dialects into Standard English.
Netflix has not yet responded to a BBC request for comment but after the outpouring of social media complaints said on Twitter that it was looking to fix some of the issues raised.
Subtitles are created in different ways by different broadcasters, with many employing outside subtitling firms. They can be written manually and time-coded to audio, or are generated using dictation software or audio recognition.
Gemma Rayner-Jones, 31, from Canterbury in England, uses subtitles to help her to concentrate when watching shows online because of a cognitive impairment.
Because she is able to hear and notice the differences, she has been tracking and complaining about inaccuracies in Netflix’s subtitles for about two years.
She estimates that she has submitted about 150 complaints in that time, and says she has not had a response.
“Everyone should be getting the same experience,” she told the BBC.
“It seems a shame to have a system to report faults there to placate people, but they don’t seem to be doing anything about it.”
She wants Netflix to be more transparent about how it handles complaints, so that users can check in whether action has been taken.
Student Chrissy Marshall, 18, studies film at the University of California and runs a YouTube account trying to raise awareness about deaf culture, accessibility and sign language.
She was one of many who took to Twitter to complain about inaccuracies in Queer Eye’s subtitling.
For her, online streaming still remains one of the best entertainment options available.
“I don’t watch cable or normal TV because captioning is always messed up or lagging. As for movie-going in theatres, the experience normally sucks,” she told the BBC.
“Netflix is what I use as a primary source for streaming because typically it is the most accessible, but even the most accessible has its issues.
“Captioning as a job is not to ‘clean up’ language, it’s to provide accessibility, full accessibility.
“We don’t care if it’s a bad word, vulgar, or maybe inappropriate, if hearing people get to know what is being said, we deserve to know as well.”
This is not an issue isolated to Netflix itself. While regulations are in place for closed captioning (user-activated) subtitles on typical television services, many on-demand services still lag behind.
One YouTube vlogger, Rikki Poynter, has dedicated years to working on accessibility on the platform, lobbying it to improve its automatic subtitle service using the hashtag #NoMoreCRAPtions.
In the US, the Federal Communications Commission has strict regulations which specify that captions “must match the spoken words in the dialogue and convey background noises and other sounds to the fullest extent possible” – but it only requires the regulations on shows on television, which means that Netflix-exclusive original series may not qualify.
The National Association of the Deaf sued and made a four-year agreement with Netflix in 2012, where it committed to ensuring all its programmes were subtitled.
Although the four-year decree has now run out, on Thursday the group told the BBC it was “disappointed that Netflix appears not to be providing captioning at the level that was promised” and said it hoped it would ensure it was using verbatim and accurate captions.
In the UK, Action on Hearing Loss has spent three years on a Subtitle It! campaign aiming to get the UK government to extend regulation to captioning of video-on-demand content.
Dr Roger Wicks, the group’s director of policy and campaigns, told the BBC that any attempt by providers to summarise or edit language on subtitles was a “very bad approach” which could lead to people who were deaf or hard-of-hearing feeling “alienated or patronised”.
“Subtitles are a replacement for speech, they’re meant to be verbatim so people have full access,” he told the BBC.
“Any attempt to summarise is offering a second-class service. I think this is well-intentioned, but it’s getting it wrong.”
He told the BBC his group intended to contact Netflix over the issue.
Mr Shannon, whose widely-shared tweets helped spark the debate, wants the company to change and check the way it subtitles its shows.
“I’d like to see more oversight on captioning agencies, more strict procedures for checking the captions,” he told the BBC.
“I’d also like to see those who are doing the hiring, such as Netflix, to check that all the files that they get are accurate, and not just assume they did everything right.
“I’m aware that it’s time consuming, but this will continue to be a problem if there are no checks and balances. Accessibility really matters.”
For many of us, meetings are a boring waste of time but technology could soon help make them more interesting and productive.
What do you do during a boring meeting? I canvassed some opinions on Twitter and the results were enlightening.
Some people compose haikus, others play meeting bingo, seeing how many pre-agreed words they can chuck in to the conversation.
Some secretly check out Grindr on their phones or watch catch-up TV, while others fiddle with their jewellery, doodle, or simply nod off.
What’s frankly worrying – if you’re the meeting holder, that is – surveys show that the vast majority of us confess to doing other things during meetings.
And there’s always one person – often a man who loves the sound of his own voice – who drones on and on so no-one else can get a word in edgeways.
Wouldn’t it be fantastic if an artificially intelligent (AI) meeting bot could tell him to shut up?
Well, that day may not be too far away.
It is “very feasible” for an AI to recognise when one person is dominating a meeting, or when a circular discussion keeps coming back to the same point, says James Campanini from videoconferencing company, BlueJeans.
“If no new points are made after a while, the AI could suggest to wrap up,” says Cynthia Rudin, a computer science professor at the Massachusetts Institute of Technology.
“While it’s a lovely idea to think everybody will be fabulous at running meetings, everybody is not,” observes Elise Keith from Lucid Meetings, a US-based meeting management platform.
An AI agent “might be able to determine whether a meeting leader is ensuring that each participant is being heard equally and fairly,” she says.
Voicera, founded in 2016 in Silicon Valley, has created an AI assistant called Eva. As well as taking notes, Eva identifies a meeting’s action items and decisions.
“If AI can do most of the mundane and drudgery work during business meetings, that leaves more space for humans to think about strategy and vision,” argues Niki Iliadis at the London-based Big Innovation Centre, an innovation hub working in AI.
In Japan earlier this year, the prefecture of Osaka – which is responsible for nine million people -started using AI to transcribe and summarise the 450 cabinet meetings it holds annually.
The AI recognises from the context whether speakers are using the Tokyo or Osaka dialects, and who is speaking as it transcribes.
So far it has halved the time needed to produce summaries and has cut staff overtime, the prefecture says.
How about not even having to be physically present at a meeting?
One feature which shouldn’t be far away is having an AI avatar join meetings for you, when you’re running late, says Mr Campanini.
So “my AI identifiable creature joins the meeting, takes notes for me, and when I join, it stops and sends me the notes,” he says.
Quite often we find we’ve been invited to a meeting that isn’t relevant to us or is at a very inconvenient time. So tech firms are also working on AIs to help decide who should attend and when the meeting should be, Ms Keith says.
One Stockholm start-up, Mentimeter, is making it easier for meeting participants to give instant anonymous feedback about whether they find a discussion useful or tedious.
“One way of solving sucky meetings is letting the audience take part in a simple way,” says Johnny Warstrom, the start-up’s chief executive.
Participants using the software can make open-ended responses or vote in multiple-choice quizzes.
When the presenter turns on the word cloud feature, a screen is updated as participants submit comments, and the most frequently used words appear largest on the screen.
Such anonymous live feedback has “fundamentally changed the dynamics of a presentation”, says Austin Broad from financial services firm AFH Wealth Management.
He now spends more time discussing unexpected responses than “simply confirming comprehension”, he says.
Mr Warstrom believes the software allows less assertive participants to have a say for once.
“All of a sudden everyone has a voice, someone at the back of the room as much as the person speaking loudest,” he says.
He thinks this is probably why Mentimeter, which has 20 million users and is Sweden’s fastest growing start-up, has more female than male customers.
But until such smart meeting tech becomes more widespread, it seems we’ll continue wasting time in the office.
According to the MIT Sloan Management Review, executives now spend 23 hours a week in meetings – up from under 10 in the 1960s.
And in one large company, a single weekly status meeting, and the preparations for it, took up 300,000 employee hours a year, the Harvard Business Review discovered.
Surveys show that the vast majority of us think they’re a waste of time. Even bosses have been increasingly critical.
Tesla boss Elon Musk, for example, told his employees in an April e-mail to “walk out of a meeting or drop off a call as soon as it is obvious you aren’t adding value.”
“It is not rude to leave, it is rude to make someone stay and waste their time,” he added.
And Amazon boss Jeff Bezos has banned PowerPoint – the bane of many meetings, particularly when speakers simply read out exactly what’s on the slides.
Many meetings duplicate work that’s already been done, so making meeting notes easily searchable could help, says Neale Martin from MeetingSense, a US-based meeting software firm.
Tools that can create agendas, send meeting invitations, distribute notes, and keep track of action items should improve effectiveness, he believes.
Otherwise, he says, “we have all this videoconferencing and other tech to link us, but we’re still doing things as we always did.”
A lot of this may sound like wishful thinking, particularly when you think how often basic tele- and video-conferencing tech fails to work.
But anything that helps meetings become slightly less painful must surely be welcomed.
Now, back to your doodling.