How AI & robotics are transforming social care, retail and the logistics industry

A closer look at what AI and robotics could mean for three sectors where a large proportion of the workforce is low-skilled

By Benedict Dellot and Fabian Wallace-Stephens

Follow Benedict and Fabian on Twitter @BenedictDel @Fabian_ws

The purpose of this granular analysis is to paint a more vivid picture of what machines are capable of, and to reveal the subtle and often counterintuitive consequences of automation that tend to be overlooked in mainstream debates.

Pepper the robot. Flickr / Collision Conf

Social care

The social care sector is home to child care workers, home-based personal care workers, mental health support officers and drug and rehabilitation therapists, among others. These occupations are often see as the most difficult for machines to automate because of the high degree of human interaction involved. Nurses in adult social care, for example, will spend much of their day persuading their patients to exercise or to eat, and will use careful communication to understand what may be troubling them. Our analysis of the OECD PIAAC database confirms these suspicions, revealing that a high proportion of the time allocated to social care roles is spent on tasks demanding social intelligence and, to some extent, manual dexterity (e.g. in lifting patients).

However, this hasn’t stopped AI and robotic engineers from seeking to automate duties performed in social care. Among the machines currently being developed are:

• Assisting robots — Humanoid robots like Pepper and Pearl can read and respond to basic human emotions. They promise to help patients by answering questions, guiding them across buildings, and encouraging them to undertake exercise. Some, like RIKEN’s Robear, are also designed to lift patients. The latest version is equipped with torque sensors and more precise actuators to allow for softer movements.
• Monitoring systems — Systems like IBM’s MERA are being deployed in homes and care homes to monitor older people and pick up on early signs of distress. AI technology is used to build up a ‘contextual understanding’ of a normal day (e.g. the times people get up and go to bed, and when they have their meals), and raise alerts with carers when anomalies arise.
• Autistic support — A specially made robot called Bandit, developed by the University of Southern California, has been used to improve the emotional development of autistic children, for example teaching them how to share through repetitive games. According to the University, the robots are less intimidating than people because they repeat their behaviour with consistency.
• Administration — In 2016, Harrow Council signed a 10 year partnership with IBM Watson to develop an AI-powered personal budget tool. It promises to ‘learn from user behaviour’ to help patients manage their care budgets more effectively and recommend more suitable support providers. It can also reportedly predict future health risks based on historic data.

These and related technologies have clear potential to substitute for workers. The transport robot TUG, which moves medicine, food and other supplies in health settings, could replace some of the work of porters. Equally, robot assistants like Pepper may one day fill in for humans in helping patients find their way around care homes. Yet we have to question whether it is desirable for social care workers to retain such roles. Moving objects around a building or reminding patients to take their medicine do not make the best use of people’s talents. Robots and AI systems may allow social care workers to become more attentive to the needs of patients and more absorbed in the core act of caring.

A good example is the use of robots in domiciliary care. With support from Innovate UK, Three Sisters Home Care in London is working with Designability, Bristol Robotics Laboratory, Shadow Robot Company, Telemetry and the Smart Homes & Buildings Association to develop a ceiling-based modular robot to be installed in the homes of its patients. Called the CHIRON project, the devices will mean that only one care worker will be required to lift a person rather than two (see The Chiron project below for more details). Three Sisters CEO, Jobeda Ali, says this will enable her workforce to do more human focused work:

“If I don’t have to send a person to do a transfer job [lifting], I can send them to have a cup of tea and a chat. This is a much better use of their time than carrying patients or cooking meals”.

Jobeda says that far from wanting to cut her workforce, the use of machines will allow her to plug vacancies and manage staff churn. “One of the reasons I got involved in this robot pilot is because I realised there just aren’t enough people to serve the needs of the care market. The use of these modular robots could effectively double the power of my workforce”. The company Skills for Care estimated in 2016 that 6.8 percent of roles in the adult social care sector were vacant, giving an average of approximately 84,000 vacancies at any one time. Jobeda believes that inventions like the one Three Sisters is building will not only help to fill these gaps, but relieve care workers of backbreaking work and make their profession more “technical in nature”.

Many of the aforementioned technologies also offer a novel function that does not duplicate the existing responsibilities of social care workers. A case in point is the phone-based app LifeGraph, which monitors vulnerable patients with mental health conditions. Its developers claim that the software can detect a mental health episode a full month before a person requires hospitalisation — not a skill that any social care worker has ever professed to have. Other inventions such as exoskeletons are complementary to social care workers. Cyberdyne, for example, has developed Hybrid Assistive Limbs (HALs) that can magnify the strength of caregivers and prevent injuries to their lumbar backs.

The CHIRON project

CHIRON is a two year project funded by Innovate UK. It strives to design care robotics for the future with a focus on dignity, independence and choice. CHIRON is a set of intelligent modular robotic systems, located in multiple positions around the home. Among its intended uses are to help people with personal hygiene tasks in the morning, get ready for the day, and support them in preparing meals in the kitchen. CHIRON’s various components can be mixed and matched to enable the customer to undertake a wide range of domestic and self-care tasks independently, or to enable a care worker to assist an increased number of customers.

The vision for CHIRON is to move from an ‘end of life’ institutional model, widely regarded as unsustainable and not fit for purpose, to a more dynamic and flexible market that offers people greater choice in the care sector when they require it.

The CHIRON project is being managed by a consortium led by Designability. The key technology partners are Bristol Robotics Laboratory and Shadow Robot Company, who have considerable expertise in conducting pioneering research and development in robotics. Award winning social enterprise care provider, Three Sisters Care will bring user-centred design to the core of the project. Smart Homes & Buildings Association will work to introduce the range of devices that will create CHIRON and make it a valuable presence in people’s homes.

Retail

Unlike care, retail is seen by many as sitting squarely in the crosshairs of new machines. According to PwC’s analysis, 44 percent of wholesale and retail jobs are at high risk of automation. Our own poll finds that 15 percent of business leaders in retail think their organisation has a high number of jobs that could be displaced in the coming decade (30 percent or more). This matters for three reasons:

1. retail is a major employer, home to 1.1 million retail assistants
2. retail jobs can be found in every corner of our country, including within low income communities; and
3. retail jobs often act as a gateway into the labour market for young people and marginalised groups such as the disabled.

Recent innovations in the sector appear to substantiate claims of impending disruption:

• Automated inventory management — AI systems are being deployed to monitor stock inventory and more accurately predict fluctuations in consumer demand. IBM’s Commerce Insights tool gives retailers real-time performance data on products, allowing them to take action to prevent over and under stocking in shops. Elsewhere, a startup called Simbe has created the Tally robot to audit shelves and spot misplaced items.
• Chatbot retail assistants — New chatbots are promoted as a means to enrich the consumer experience. The North Face’s customers can search for items on their website through natural conversation with an AI interface. Similar innovations are appearing in physical stores. MindMeld provides conversational tools to retailers include Uniqlo, allowing customers to ask questions they typically would of shop assistants, such as the whereabouts of items.
• Enhanced search engines — Excitement is growing among retailers about image-based search engines underpinned by deep learning algorithms. Rather than writing terms that describe a product, customers can take a picture of an image and the algorithms will find matching items. Asos, John Lewis and Nordstrom are all reportedly developing apps or website functionality to power visual search.
• Automated e-commerce design — Marketers will be familiar with A/B testing to determine the hit rates of different website designs and brand messages. AI promises to turbocharge this process through evolutionary algorithms that continually test and refine content using feedback data. The Italian lingerie company, Cosabella, worked with the startup Sentient to rapidly test alternatives for its website, resulting in a sales uplift.

The impact of these technologies could be severe. Self-service checkouts may eventually rid many cashiers of jobs, while inventory management systems might one day eradicate the need for store managers. However, there are several reasons to doubt such fears. Number one is that some technologies are still a way from delivering on expectations. One senior call centre director writes: “The day when chatbots handle all customer queries and contact centre agents are completely replaced is a very far off day”. Similarly, a recent investigation into self-service checkouts found they took consumers longer on average to make a purchase than going through a traditional kiosk, and continue to be undermined by shoplifters.

Another factor is that some of this technology will augment workers rather than displace them, as we saw with social care. On the marketing side of retail, new AI systems are emerging to help sales teams find leads within client companies, and then to tailor messages so they are more likely to resonate. LeadGenius can pick out top decision-makers with buying responsibility across a range of companies. Another startup, Chorous, can analyse the content of sales calls to highlight topics that repeatedly crop up, or particular pain points that are emphasised. In some bricks and mortar stores, AI systems are being used by employees to answer customer queries more accurately, rather than have customers directly interact with chatbots. Holm is an application designed to be used by assistants so they can better match customers with clothing items.

What about the potential for AI and robotics to shift trade from retail stores to online? If innovations such as image search engines and sophisticated recommendations take off, this could move more spending online and take away jobs from the high street. Again, however, this fear deserves closer inspection. Evidence from both the US and the UK indicates that job losses in retail may be compensated for by job gains in the warehouses that underpin e-commerce. The US economist Michael Mandel estimates that the e-commerce sector in the US has created 355,000 new jobs since the crash, compared with the 50,000 lost in retail. Our analysis of the UK shows that since 2010, jobs in the retail trade have fallen by 7,000, while jobs in warehousing have increased by 115,000. The latter jobs are also better paid on average (see below).

Much also depends on the business strategies of retail companies. To the extent that supermarkets, clothing retailers, electronics stores and other outlets prioritise cost savings, we can expect many retail jobs to be lost to AI and robotics. Yet some retailers will pursue a strategy centred on customer experience, with greater emphasis given to person-to-person interaction. A good example is Apple’s retail stores, which have multiple staff on hand to create a rich experience for customers. While research on the strategies of UK retail businesses is limited, a recent US study by Cornerstone Capital Group found that no retailers in its analysis pursued a ‘convenience’ strategy predominantly based on cost, while 35 percent positioned towards an ‘experience strategy’.

Logistics

The logistics industry encompasses the organisation, storage and movement of goods across supply chains. Even more so than retail, logistics is thought to be on the verge of an automation revolution. Approximately 40,000 robot units were shipped to warehouse and logistics businesses worldwide in 2016 — a figure that is expected to jump to 620,000 by 2021. Leading the way are companies like Amazon, which is continuously on the lookout for efficiency savings in its sprawling supply chain. Logistics businesses are also being pushed to innovate by consumers who desire ever faster and cheaper (if not free) delivery, and who want items to be shipped at a time of their convenience. Several AI and robotic systems stand out as potentially game-changing for this industry:

• Warehouse robots — In 2012, Amazon bought robot maker Kiva Systems for $775 million. Its bright orange robots shuttle pallets and product shelving units around warehouses, allowing workers to pick and pack goods without moving through the aisles themselves. Other brands of warehouse robots have since emerged to aid order fulfilment, including Fetch Robotics, Locus Robotics, and Vecna. Many are powered by LiDAR sensor technology, which uses light and radar for navigation.
• Supply chain management — Shipping companies must manage extremely complex distribution networks, making sure that goods arrive precisely where they need to while giving an appropriate freight price to wholesalers and retailers. IBM’s Watson Supply Chain uses artificial intelligence to determine the optimum route for cargo by crunching live and historic data on weather patterns, port congestion and natural disasters.
• Anticipatory logistics — Anticipatory logistics refers to the process of predicting demand for consumer goods before purchases have been made. This allows logistics firms to improve efficiency and cut delivery times. Ocado, for example, uses algorithms to optimise its warehouse storage structure, meaning popular and soon-to-be popular items are in plentiful supply and in close proximity to its picking and packing teams.
• Self-driving vehicles — Several logistics and technology firms are trialling autonomous vehicles for goods delivery. Ocado recently worked with startup firm Oxbotica to pilot the delivery of groceries in London via a driverless truck called CargoPod. At the haulage end of logistics, Uber and Starsky Robotics are both developing systems for managing autonomous HGVs.

At first blush, the potential for jobs to be displaced in the logistics sector seems considerable. Yet just as with the social care and retail industries, what appear to be impressive technologies on paper are often incomplete and have limited functionality. For example, despite feverish attempts to create gripping robots that can pick up items and stow them neatly into boxes, no machine can yet match the dexterity of humans to do so. The last Amazon Robotics Challenge event, which brought together robotic engineers to compete on a gripping robot challenge, revealed that even the most advanced machines continue to have difficult handling items that are wrapped in plastic, obscured, or which bend and change shape when moved.

Humans are also likely to remain in place for ‘last mile’ delivery in logistics (i.e. getting packages to the doors of customers). An innovation manager for a major logistics firm in the UK told us that robots would be ill equipped to deliver to gated buildings and high rise apartment blocks — recall that machines can find the simple task of opening a door difficult — or where some interaction has to be made with recipients (e.g. helping them to install items or finding a neighbour to take in a package). Regarding the prospect of delivery by drone, the innovation manager felt this would be extremely difficult in crowded urban areas, although a possibility for rural locations. He said plans for drone delivery and other zany schemes are more likely to be PR ploys than game-changing innovations with commercial potential.

The logistics sector will also see the emergence of machines that collaborate with human trainers and operators. US-based Starsky Robotics is a firm that creates technology to power self-driving trucks, but with a clearly defined role for humans to step in at key moments. Their vision is for HGVs to run on long stretches of highway unaided, and for remote drivers to take the reins of the vehicles in the final furlong of delivery. The company, which has raised $3.75m from Y Combinator and other investors, expects each office-based driver to monitor and control between 10 and 30 trucks at a time. If a technology like this were to take off, it would ultimately lead to fewer HGV drivers but arguably better quality jobs.

This article is an extract from the RSA report The Age of Automation: Artificial intelligence, robotics and the future of low-skilled work. For full references and bibliography please visit the RSA website to download the full report.

To find out more about our research, please contact Benedict Dellot