The impact of AI in a transformed world
Artificial intelligence has entered the mainstream and the fourth industrial revolution is continuing apace, leading sceptics to question what this means for the average employee.
As we look towards the re-opening of the economy, it is clear that AI offers huge potential for a new world of work that is safer, more productive and more innovative than ever before. A medium- if not long-term effect of COVID-19 is an immediate need to reduce workplace capacity, and an acceleration in automation that will increase the safe space for remaining employees to go about their day.
With social distancing, remote working and the risk of future peaks leading to a potential increase in sick workers, AI will be vital in enabling new ways for businesses to operate.
Some companies are ahead of the curve; almost two years ago, online grocery service Ocado’s revenue saw an increase of almost 12% thanks to the automation of its warehouses. With this in place, the company can process more than a quarter of a million orders each week.
For healthcare, AI also has the potential to be vital, whether in delivery of care – from disease detection to treatment and even surgery, as well as in helping with hospital efficiency and workflow, which itself can enhance care delivery and cost of treatment. Human error is estimated to be around 30% in high-pressure decision-making scenarios, a risk that can be reduced exponentially by integrating machine learning and AI into the process.
Automation also offers the indirect advantage to the supply chain of more accurate forecasting of demand. Food supply for example can therefore be redirected from closed/ reduced capacity cafés and restaurants, and towards over-burdened supermarkets, preventing inessential contact along the supply chain. This will be reduced further through advances in autonomous vehicles and drone delivery, which will become ever more important.
Not only will this be important for manufacturing, but in the mobility sector. Social distancing and fears over future peaks of the virus mean many are looking to avoid public transport. An acceleration in the wide implementation of mobility solutions such as autonomous vehicles could provide an alternative route to work for many, that does not risk exposing them to drivers and other passengers.
The benefits of AI have been discussed at the annual CogX festival for the past three years. As we head full throttle towards our fourth, and first virtual, edition of the festival next week in the midst of this pandemic and rising societal unrest, the necessity of it is becoming ever clearer. But of course, we have to equally acknowledge the very real risks around this technology and the ethical questions AI continues to present.
Perhaps most pertinently, with the rise in activism around the Black Lives Matter movement following the death of George Floyd in the United States, the very first concern to address is how we can make artificial intelligence that eliminates racial bias. The technology industry remains disappointingly and frustratingly un-diverse, and this has heavily impacted on AI’s development – as machines learn from the environments they are built in, resulting in them learning the same biases and prejudices of those creating them.
The examples of how these biases has translated in reality have varied from the baffling to the truly grim – from automated soap dispensers that would only release soap onto white hands, to driverless cars that do not recognise black pedestrians, rendering more likely to run them over.
This is an urgent problem for technologists to find a solution for. Innovation must be to the benefit of everyone, and the protests in the US right now serve as only the latest reminder than we as an industry must do better in representing black people.
It is not only black communities at risk of exclusion. Israeli historian and author, Yuval Noah Harari argues in his second book, Homo Deus, that the rise of artificial intelligence will lead to a new ‘useless class’ of humans: with technology accelerating so quickly that skills will go out of date faster than people are learning them, leaving many unable to work.
It is, even by Harari’s own admission, a worst-case scenario. But it is a legitimate fear: how do we ensure continued technological advancement does not compromise people’s ability to contribute to the economy? The World Economic Forum predicts that by 2025, more than half of all jobs that existed in 2018 will have been replaced by automation and before 2030, people will need more than 100 days of learning each year to keep up.
This concern has only been accelerated by COVID-19; the United Nations’ labour body estimates that globally, 195 million jobs will be wiped out. Not only have hundreds of millions lost their jobs in the short-term, but there are concerns in some industries that these roles may never return. If AI takes their place, then these workers will urgently need to be upskilled to take on replacement roles, or risk becoming the first casualties of the useless class.
However, there is hope that emerges from these questions: every industrial revolution leads to a period of re-skilling and adaptation, and each time the jobs that humans then take over are safer than those of previous eras.
A study in The Lancet argues that a personalised prediction machine learning model can be used to create safer work spaces and communities by measuring individuals’ vulnerability and risk of suffering severe outcomes to diseases such as COVID-19. These models would help to allocate treatment more effectively, and help in preventing future lockdowns and creating safe work spaces by not unnecessarily isolating the least vulnerable in society and shielding the most.
Within work spaces, thermal imagery and temperature checks are likely to become commonplace as businesses look to reduce the risk of employees and customers with mild or even no symptoms entering a public place of work. To reduce the intrusiveness of these measures, we can expect smart cameras to be increasingly installed at the entrance points to office buildings and stores around the world, automatically detecting any possible symptoms
Where the biggest ethical questions around AI remain – and we will be debating this at CogX next week – is over privacy. Author Shoshana Zuboff theorises in The Age of Surveillance Capitalism that the next stage of market capitalism is the exploitation of the human experience for behavioural data. Those operating businesses that use these algorithms are effectively betting on consumers’ future behaviour based on existing data and are using it to become wealthy.
Ethical questions arise around how conscious we all are of the data we give away, the power we have over our own data and our ‘right to be forgotten’ due to the General Data Protection Regulation of 2018. Every single time we give away our data we are trusting a company to use it in a way that benefits us, not exploits us. If the companies who own that data use it in illicit ways, the trust between consumer and business will be eradicated and people’s personal freedoms will be compromised.
Next week, CogX will be focused on how we get the next ten years right and through our virtual festival, is bringing together leaders, CEOs, entrepreneurs, policy makers, artists, academics to look at this in the context of so many global issues. AI will be a crucial part of this – with a role in ensuring the economy gets moving in a way that protects people’s health, inclusion for everyone in society, and for the future of work.
There are ethical considerations and natural risks and consequences to an industrial revolution of this scale that go alongside this, but this should not mean we run away from it: to get the next ten years right, we have to get AI right, and to do this we must embrace it and work collaboratively to make it work for everyone.
Charlie Muirhead is the CEO and founder of CogX