From climate change through refugee and pandemic crises to the rise of populism and conspiracy theories, the 2020s can be an age of anxiety played out in the glare of 24/7 media on camera and online.
In the middle of it all, we keep on keeping on. These are stories of lives lived today. There is love and laughter, big decisions, hope and reflection on what matters most.
From climate change through refugee and pandemic crises to the rise of populism and conspiracy theories, the 2020s can be an age of anxiety played out in the glare of 24/7 media on camera and online.
In the middle of it all, we keep on keeping on. These are stories of lives lived today. There is love and laughter, big decisions, hope and reflection on what matters most.
The presenter stood at the podium as the applause died down. He looked seriously at the audience.
“And now ladies and gentlemen, a special award. We all realise of course that the huge achievements of AI start with machines learning to make decisions and that they are taught how to do that by humans. Our industry is therefore indebted to an army of often unsung heroes who make it happen. I am talking of course about the crowd workers whose dedication enables the machines to learn. It therefore gives me great pleasure to announce that the special award goes to…………….”
As Vidya heard her name, she involuntarily cupped her two hands over her mouth. It was a classic Oscar ‘nominees at the table’ shot that the media would re-run numerous times to emphasise her humbleness and indeed maybe the humbleness of all the workers that she represented.
As she walked to the stage amidst the applause, Vidya was not thinking about the words that she would shortly speak. She had not imagined ever being recognised let alone being here so winning was even farther from her mind. There were however a number of important issues that needed to be raised and, despite the emotion of the moment, she drew on them as she climbed the steps to the podium.
“I am…I don’t know what to say. Really…I.” The sentence tailed off as Vidya composed herself. The audience recognised the pause. “Thank you so much for this award. It goes without saying that I never expected it or even dreamed of it. For my family and friends who also work in the back rooms that support the development of AI, it is a great thing to be recognised. This award is definitely for those many many people in my country and beyond. When I first started picking up tasks on the computer in my apartment in Hyderabad, I was not thinking about the bigger picture – about how my human responses would help a machine to perform tasks that it would never otherwise understand. As I began to recognise that we were teaching the machines, and the machines were changing the way the world functions, it was sometimes quite bewildering.”
In the audience, a number of Executives smiled at one another. It may have been out of pleasure at the achievements of Vidya, but it was more likely at the reference to their own work revolutionising a worldview.
“As I categorise news stories and videos or I translate statements and questions or add captions to videos, I know that these small tasks have to be done by us – humans – first today. Only then will such decisions go towards enabling machines to make them tomorrow. By being here to accept this award I am happy to reflect your generous acknowledgement of our role in that transition.”
Applause rolled around the room.
For the first time, Vidya allowed herself to look up and out at the audience. The stage lights in her eyes obscured any faces but the wealth and relative youth of what was an overwhelmingly American and European group was very clear. She was again struck by the feeling – as she had been since arriving – that all this glitz and glamour was a long way from Hyderabad.
“I know that I am lucky to be a part of this great industry. After graduating with my degree, I took some time out after our baby arrived. Returning to the workplace at the age of 31 was not easy in my country. I was so grateful to find a role that enables me to work from home, fit my work around my family, hone my language skills and develop new skills too. I am blessed.” Vidya took a second’s pause and several in the audience thought that the moment’s emotion warranted the same for themselves.
The presenter lightly touched Vidya on the shoulder and when she looked at him his expression was of reassurance – it said, ‘take your time; there is no hurry.’ She was, however, approaching a feeling of being overwhelmed. That sense of her not belonging had never left. Indeed, it seemed to grow as the speech progressed. She wanted it over and felt that those last words were really the only ones necessary.
“Thank you all so very much.” She stepped back and the room exploded into a crescendo of applause. As she stepped down from the stage and made her way back to the table, a sea of people rose to their feet and turned to offer their congratulations and smiles. It was a far more personal show of appreciation than she could ever have felt from the stage and was all the more moving because of it. As she retook her seat, her composure was close to breaking point and she longed for the applause to subside so that she could once again return to being a face in the crowd as the lights went down.
It was only later, in the early hours of the morning, when she was back in the quiet of her hotel room that she could luxuriate in the warmth of recognition and appreciation. It was genuine and affirming. That work could be so much more than paying the bills – that it could be a part of who she had become and would yet become – was everything to her.
***
None of that ever happened, of course; nor is it ever likely to. It was the product of a daydream. Not by Vidya, who had no time for such things, but by someone over 4500 miles away on the other side of the world.
Elliot had time for such things and he had reason for them, too.
Life had changed so much in a matter of weeks. After three years at a Nottingham software company as a Product Manager, Elliot had handed in his notice. He had an exciting job offer that would pay him more money and give him more responsibility. It was a dream move. A week before his months’ notice ended, however, the Covid-19 pandemic lockdown hit and the job offer was withdrawn. He was left high and dry – his old job gone, his new one no longer there and no qualification for furlough.
At the age of 30, Elliot found himself dipping into his modest savings and realising that they would not go far. He’d always seen the importance of having at least something behind him to hedge against a rainy day but now it was pouring. There was some help at hand – a mortgage holiday, for instance – but it represented just a small chink of light and as the first lockdown transitioned into the second and third, a sense of desperation took hold at times. Friends and family (especially his parents) offered a helping hand, but he knew that he had to do something himself.
Since his specific work qualifications and experience counted for little when there were no relevant jobs out there, he looked beyond to pretty much anything. Covid was its own growth industry of course so he applied for Track and Trace Contact Tracer roles, Testers, Administrators, Co-Ordinators and – later – Vaccinators.
He also looked at part-time work, although there was little about it that appealed – inflexible and unpredictable hours that nevertheless required his commitment, poor money and no benefits. In any event, little of this had survived the pandemic anyway.
He reluctantly but inevitably ended up at the so-called gig economy, now no longer just a media report or pub conversation but a very personal prospect. He needed something that could be done from home and in his own time because he had to continue looking for work and would want the flexibility to immediately switch if anything came of the applications.
That’s how he found the crowd-working platforms. Or rather, it’s how he noticed them. In fact, the development team at his old company often referred to tasks being put out to crowd working but he’d never before taken much notice. Now it had become personal.
He recalled software design meetings in which they discussed on-the-fly decisions requiring a level of subjectivity that machines couldn’t handle. From virtual assistants to product searches, there were untold numbers of cases in which they could not design a yes/no answer; a binary machine decision wasn’t possible so they needed humans to create answers that would become ‘training data’ to teach the machine how to do it next time. Elliot had never pursued just where the developers had acquired this training data from. He remembered them saying that the budget necessary was modest and that it could be generated very quickly too, so no one gave it too much thought beyond that – the machine simply needed to be fed.
He distinctly remembered the term that was used, too – ‘the last mile.’ This was the element outside of the skilled computer scientists and visionaries. At the time he’d thought it would better be called the ‘first mile’ because it was the essential starting point. Nothing happened without it. But this wasn’t Elliot’s role and he didn’t have time for idle curiosity.
Now, though, he was on the other side and time was all he did have. He needed to understand crowd working. He first came to understand that crowd-work platforms sat between the company requesting a task and the workers who completed it. There was no direct relationship between the two, so the platforms held huge power in the middle. It would be a very important lesson.
At first, navigating the process was bewildering. He looked around at the type of tasks out there and the rate of pay. It was clear pretty quickly that the range was huge – work was broken down into tiny packages of repetitive tasks that were consequently deemed unskilled. The terms were generally pretty homogeneous – low unit rates of pay, rigid processes to collect that pay, fast turnaround requirements and high standards that had to be met to avoid negative scores blocking future work.
Most of all, there was absolutely no commitment to the worker from the requestor companies or the crowd-work platforms. If he went this route, Elliot would (like all gig economy workers) be self-employed and on his own. No sick pay, no holidays, no pension, no benefits and no contracts.
Then he (virtually) met Vidya. He’d later look back on his good fortune at discovering a friendly presence in such a cold and anonymous landscape. By then he had registered with three or four crowd-working platforms so these were early days and he had a long list of concerns and questions.
Finding one or two independent forums for crowd workers, Elliot started to post queries and receive the simple and straightforward answers that the platforms themselves avoided. Noting the time differences and some of the user names, it was pretty obvious that many of these came from the Indian subcontinent. At times he was embarrassed by the naïveté of his own questions but he started to recognise in the answers a spirit of self-protection for this far-flung community of which he was a hesitant and alien new member.
As helpful as the answers were in general, Vidya’s went farther. In their exchanges, Elliot felt he better understood her and that he in turn opened up more of himself. By his standards, she was a veteran of crowd working and had already seen off a number of obstacles that he himself would not encounter. These included scam platforms and those charging fraudulent fees to register with real ones as well as the worst paid tasks (the platforms only posted tasks to the countries where the rates were realistic; Brits would not accept some of the rates that Vidya and her friends had to.) She had a well-tuned process for finding the best tasks and delivering them efficiently.
“Don’t consider that one,” was one of the most valuable posts. The brevity was helpful. Elliot had learned it quickly and he noticed that more often than not it came from Vidya. Sometimes these would be followed by a short explanation like “unclear instructions so good chance of rejection” or “bad payer.” Sometimes she’d point out that some of the best tasks were limited (“you’ll only see this if you have a high score over time”) which gave him a glimpse of some of the promise that she might well have kept to herself.
Sometimes the advice was more personal. For a man with a history of working to tight deadlines, Elliot was surprised at the impact on him of the appearance of a countdown timer on the screen. Everything was measured to the nearest second. It was a constant reminder that the task had to be done quickly otherwise it would be passed on and he would not be paid. Vidya advised ways to zone out of that tension and instead concentrate on the task itself. Nothing seemed to faze her whereas, at times, everything fazed him.
“That can be very disturbing,” was another guidance comment. Vidya’s understatement was obvious when Elliot realised that the task was defining images as pornographic or violent. The cesspit of the internet is diluted by the removal of these, but machines can only spot what needs to be policed once they are trained to recognise it. The human role here was amongst the toughest and Vidya had, through experience, learned how to avoid these wherever possible. When she’d started, there was no way to do so and sometimes it would still be necessary if they were the only available tasks paying what was needed.
The sheer number and range of tasks was an eye-opener for Elliot, as was the number of people necessary to do them. The requirements were tough and it became increasingly clear that the platforms worked on a ‘take it or leave it’ basis. Once in a while he looked at it from the perspective of his old job – it made sense when you have an army of low-cost workers to set the bar high as there will always be another along to replace anyone who can’t (or chooses not to) handle the task. It felt different though when he was that worker.
As a worker too, he was stunned at the minuscule payments. Page upon page of tasks paying pennies. There was no way on earth that this could ever be more than pin money. It was no solution to his dilemma and yet he had to do something. He had no choice. He had to find a way to navigate around the pitfalls and make the best of whatever was there.
“I try to concentrate on tasks that will improve my English, translation ability and computer skills,” Vidya had said in an exchange on how she chose tasks. The chat made clear that, for Vidya and many of her friends and family, this was not just about the money. They may have been desperate but they were every bit as determined to use it as route out of desperation. They used the tasks to develop skills that would improve their chances of better – and proper – employment.
At times when Elliot felt sorry for himself and looked enviously at his friends whose careers were simply on pause, he was shamed by the fact that Vidya took these ‘mundane’ tasks and used them (without complaint) as a ladder. For her, they transformed the mundanity into a necessary lifeline to something better. He merely had to hang in there.
By then they had properly identified and introduced themselves, at least as far as anyone does in a forum. They had stepped partially out from their IDs and usernames.
Vidya, like many others on the forums, was well-qualified and very intelligent. Indeed, on the few occasions when people spoke about their educational background, it became clear that the vast majority were much better qualified than Elliot. His daydream of Vidya’s award portrayed a stereotypical view of these invisible ghosts but it was wrong; their expertise and value were simply not recognised or not buffed up with the corporate sheen of the companies that benefited from their hidden work.
Vidya’s husband, Adhik, worked full time and was originally reluctant for Vidya to work. He knew however, because of the massive gender imbalance in full-time employment, she had to make something happen for herself. Now that she had, he saw the value not only in the income but in the pride that his wife took in her achievements and the possibilities for their future. Even via the coldness of an online chat function, Elliot thought he recognised the warmth in Vidya’s telling of the story. His own singledom was a positive in these dark days but he saw too in Vidya’s and Adhik’s story a sense of how togetherness helped pull them through – and up.
Elliot would feel guilty when he realised that he was exchanging messages with Vidya after 6 in the evening in Nottingham which was approaching midnight in Hyderabad. That time shift was a daily normality for Indian workers whose body clocks were necessarily set to the times when tasks would be posted – European or (much worse) US time. He was only too aware also of Vidya’s other family tasks – not least caring for their son Vitasta – especially during those July and August days of scorching heat and pouring rain.
Vidya and Adhik lived in a one-bedroomed apartment outside the city centre. Adhik travelled into the city each day – he would have liked to live there but their combined income didn’t allow it. As they would need more room soon for Vitasta to grow, their combined focus was on working and saving. Vidya in particular knew that she could increase her income by taking on more tasks and organising her days around them.
She had become single-mindedly focused, always with a view to getting ahead rather than just surviving even if that meant taking on the worst tasks over the longest hours. In a sense, this meant taking a step backward to move forward. When she first started crowd working she took on anything and everything – the worst and most disturbing type of image labelling, the most offensive hate speech, the most robotic and repetitive tasks at any hour of the day or night. She invested her time solely in creating high scores for herself. Those were tough times. Adhik was furious at some of the images that she was seeing – even though he knew the reasons – and worried at the impact on her health of her never having any rest. Several times too the platforms would just go down – or she would be locked out from her account – and the work would dry up; any joy they might have taken from the respite from work was more than overshadowed by the sudden and uncontrollable loss of income. Those were dark distant days that neither wanted to return to. For their future however, Vidya would probably need to – at least in bursts – if they were to go beyond mere survival.
For his part, survival was exactly what Elliot was looking at. He knew, however, that his survival would likely be her luxury. That never left him.
“I have to go now to my mother,” would be the phrase that Vidya most often used to end an exchange. Elliot didn’t know what, if any, responsibility she and Adhik had for their parents but it was ingrained in the conversation. The subject was too personal for him to ask but it was always there as evidence of another aspect of Vidya’s diligence and urgency.
For Elliot, his responsibilities were for himself. Whilst – like everyone – he worked primarily for money, he missed more than that now. Being cast adrift meant that he had kept in touch with just a few people from his last job and contact with them – as with his friends – was now fading fast or restricted to the odd remote message.
The depth of Vidya’s family and friendship relationships (which were typically Indian) had served to underline the shallowness of Elliot’s isolated – and hardly atypical – British life. He knew that the forums through which he had found Vidya had some social role too but there he felt himself a fraud, an interloper into a world of conversations in which he didn’t belong about pressures and hardships that he couldn’t understand. In an odd and illogical way, this seemed to deepen his feelings of isolation.
And yet it was Vidya’s resolve and determination that most lifted Elliot’s mood. He hesitated to describe it as inspiring because of that word’s chronic overuse. To him it meant not just the admiration of something or someone but about changing your own behaviour because of it. The word was common but the action was rare. Here, though – in Vidya – it was appropriate. He felt changed by knowing her, even though it was a shallow and remote knowledge.
He wanted to do something.
***
It happened completely out of the blue. A call from someone that Elliot had previously worked with and an invitation to interview for a temporary project that, rather than being delayed by Covid, was actually being accelerated by it. What then took shape was one of those happy and lucky coincidences in which what the employer wanted was exactly what the candidate offered. So it was done. Elliot was out of the crowd-working crowd and back into a work life that – apart from lockdown – looked familiar.
The experience and kindness of Vidya never left him, though. The new project included the development of tasks for which machine learning was necessary and, though Elliot had no need to get involved in the sourcing of training data, he made it his business to do so.
When workers signed up for tasks, he ensured that instructions were thorough and that they included guidance on how the company expected the results to be represented and as well as an outline of how they would be used. This was far more than the rigid output and included more about the culture of the company – it was something that Vidya said was always absent but it made the task much clearer, less prone to misunderstanding, and it gave the workers a small glimpse of the organisation for which they were indirectly working.
He ensured too that there was a direct feedback channel so that workers could ask questions and raise queries outside of the platforms. This was hugely unpopular with his new colleagues and some fought against it, but he insisted. He never wanted workers to be left in limbo waiting for the platform to respond to queries that got in the way of a task being done and workers being paid. He needed to remove the middle man.
He set up a forum for workers to share feedback on tasks and was delighted when some of the comments were taken up by his colleagues as suggested improvements. He reported up the line the value of these suggestions and their source. When there was no one available to monitor the forum and manage responses, he did it himself – often well into the night. This took him much closer to the reality – much closer to workers like Vidya, her friends and family.
None of this shifted the feeling he had that no one understood – let alone appreciated – what happened in the shadows. This needed fixing.
“It’s just a resource,” had been the response from several of his new colleagues when he first started talking about how crowd workers were being used for training data. He understood their point but he didn’t get why no one was interested in giving it any further thought. He was often met with a curt response.
“Just leave it,” he had been told by people baffled as to why gave it so much – or indeed any – attention. They questioned how he had the time to be meddling in things that weren’t his concern. At those moments, he recognised that he should back away and bide his time, hoping that the right moment would arrive.
It did. He needed an audience of the great and the good and he found an AI event that gave him just that. The rest was about his own commitment to go beyond concern or even outrage and to act.
When he booked his place, he neglected to mention that his rather impressive job title was temporary and that he was a contractor rather than an employee of the internationally known company next to his job title. The inflated sense of importance conferred on him would secure him the seat that he needed.
The event was focused on celebrating the achievements of AI and laying out its future. Speakers included senior figures from the biggest companies developing AI and those delivering services using it. It was one of the first events to take place after travel restrictions had been lifted and so delegates came from far and wide, keen once again to mix, meet and back slap. All in all it was perfect.
Unsurprisingly, the agenda did not include any reference to the ‘last mile’ and certainly not its role as a new shape of exploitation. Elliot decided to add it.
He arrived as early as possible in order to ensure that he claimed a seat in the front row. As the lights dimmed and the chatter subsided, the event host took to the stage and welcomed the delegates before introducing the first speaker – an ‘elder’ of the industry whose opening remarks would, it was promised, provide context for the day.
“In my opinion and experience, there has never been a technology that has developed as fast or that will have as much impact on society as AI.” It was a typically huge and portentous opening statement and one that the delegates naturally welcomed. It confirmed that they were in the right place at the right time; that they were the exulted innovators and drivers of change.
Several presentations championed the view that, contrary to popular fears, AI would not create mass joblessness but instead usher in mass redeployment that would likely provide more fulfilling work or more leisure time for workers rescued from the drudgery of repetitive tasks. It was a suitably utopian vision but Elliot became increasingly more convinced that there was a strange myopia at play. The big human benefits claimed for an AI future seemed to ignore the human costs of its present.
The question for Elliot remained when to make his move. There were questions at the end of each presentation but they were often very rushed. The Question-and-Answer session with a panel of experts at the beginning of the afternoon session was longer – it would cover both what had been spoken about in the morning and what would be in the afternoon. This looked right. Too often, he had seen the agenda for these sessions set by the host so it was important that Elliot was amongst the first in the queue to pose questions that would enable him to influence it. By being in the front row and ensuring that his hand was one of the first raised would, he hoped, give him the best chance.
When the smartly dressed woman from the event organiser passed him the microphone, he was only the second speaker so he felt confident that he would have the time to make his points – if he were allowed to finish. His pulse was racing and his mind for a split second struggled to focus but as soon as he had spoken his name and company, the words – previously well practised – seemed to flow.
“Several speakers this morning talked about AI meaning machines and humans working together rather than machines replacing humans. I agree and we can see an example of this today in the last mile where humans play a key role in creating training data on which AI applications are built. Is this not however a worrying example of how unbalanced this relationship can be? So much of this work is done by invisible distant pieceworkers for miniscule payments and without any rights. Is that the foundation on which the industry should be built?”
The host was visibly uncomfortable and there was an unusually long pause in his offering the question to the panellists for response. Fortunately for him, one of the morning speakers stepped forward.
“Well, we certainly recognise the importance of training data as our raw material. It is, however, subject to the appropriate commercial terms of its environment just like any other service. I don’t think that we can really comment on that as it’s outside of our control.”
“Is it?” countered Elliot. “There are only a handful of crowd-working platforms and practically all of the major players who develop AI use them so I think it’s very much in our control.” The host appeared about to intervene so Elliot knew that he had to persist. “The crowd workers who are the foundation of the industry are entirely at the behest of these platforms. They can be locked out of their accounts or be suspended with no explanation and no recourse. They’re unable to earn whilst this is going on. It is clear that these are not the workers – the talent – that you have been talking about today. Our brave new world is based on the very visible creative talent of the brightest data scientists. Don’t you think that we should at the very least be recognising – but preferably also supporting – the invisible talent too?”
“We are getting a little off topic here so I thank you for your question and let’s move on,” said the host as he nodded to the woman in the audience to retrieve the microphone.
“It’s not off topic at all,” said Elliot. “If we are here to talk about the impact of AI on the workplace, why should that exclude the very place where much AI starts?” The host now looked more urgently at the woman in the audience who advanced on Elliot, but his grip on the microphone tightened. “The average UK FTSE CEO earns 133 times more than the average worker – if the calculation were done on crowd workers it would be many multiples of that. Since the great and the good of AI are in that category or heading there, it seems a very relevant question.”
The host was now looking to the side of the hall where security men stood and was clearly wanting them to take over from the woman in retrieving the microphone. He did not, however, ask them to and the momentary hesitation allowed one of the speakers on the platform to comment.
“I don’t think that that comparison is particularly meaningful,” he said. Before Elliot could speak, he continued. “But I take your point. Training data and the last mile are maybe both a little too opaque to many in the room and in the industry. I for one will certainly commit to look at it from our company’s perspective.” It looked like a positive commitment and maybe therefore a win for Elliot. It might, however, have been a fop to end the debate and move on.
“Thank you,” said Elliot. “I’d just like to ask one related—”
“—No, I really must insist that we move on to someone else now pl—” interjected the host.
“—This morning we talked about using AI to create fairness and to remove discrimination. We also—”
“—Please no. Let someone else sp—”
“—talked about how AI must be used ethically and all prosperity created by it must be shared evenly. Surely that should start with how AI is built because that is right now. It’s easy to talk about the next generation but we can do something about this one and—”
“—No more.” The host now stood and was almost shouting as Elliot stepped from his seat up onto the stage. His front-row position made it easy and quick. At that moment, the security men – previously only peripherally involved in the eyeline of the host – stepped forward and walked hurriedly towards the stage.
“I’m one of you,” Elliot shouted to the audience, no longer addressing the panel. “This is about what all of us stand for. That’s all.” The security men took him by the arms and walked him to the side of the stage closest to the exit. There was no struggle. Elliot had said what he had to say.
In the lobby outside the conference room, he could not hear what was being said although the hubbub died down quickly and there were a couple of loud laughs which Elliot took to be the host attempting to make light of what had happened – probably involving some insult at Elliot’s state of mind. He didn’t care. There was an odd and illogical calm passing over him.
After a few minutes, he had gathered his thoughts and readied himself to leave. There was no doubt whatsoever that today would damage – maybe permanently – his job prospects but he was enveloped in an irrational euphoria in having said his piece. What difference it could make was another question.
At that moment it would have been wise to have considered his own future but all he could think of was Vidya.
This collection of short stories by Nick Fuller does a pretty good job of capturing what it's like to live in 2020 AD - not just in terms of living through a global pandemic, but about generational fears over climate change and global warming, about generational gaps between Gen X and Gen Y, about shifting cultures, and even about technology, particularly in terms of AI and machine learning.
One of the stories I really enjoyed was the first one, "Out of the Shadows". The story is a critique of current perceptions of machine learning, particularly the work that goes into it. There are thousands of people who work thanklessly to provide data for machines and AI to learn, creating these massive data sets that will make these machines far more viable. Instead, machine learning is hailed as a project that's achieved by an elite few, ignoring the efforts put in by crowdsourcing altogether. It was a very interesting perspective that no one really talks about. Another one, "Discovering Home", isn't just about the cultural shifts between America and England, but about personal ethics in today's business world, and how cut-throat corporations can be. When it comes to making the right choice, do you go with what benefits your client, or your company? If you cross your company, would you be lucky enough to land on your feet?
Each story provides a different insight, a different peek into the various perspectives that make up what it's like to be alive in 2020 - whether it's as a corporate worker, as a footballer, as a man who's lost everything, or as someone who knows that their children will see the world burn, and can do nothing to stop it.
Nick Fuller's writing is succinct yet heartfelt, and his storytelling abilities range from short, revealing snippets to longer tales that flesh out deeper meaning, and prompts questions from the reader along the way. These Days is a pretty inclusive collection of short stories, one that remembers to capture different perspectives of what it's like to live through, well, these days.