A jet aircraft for the mind
Of the many stories about Steve Jobs at Apple, one that resonates with the authors is his early ‘80s description of computers as a ‘bicycle for the mind’. Jobs had seen a chart showing that many animals were much faster and more efficient moving over land than humans, but that a human on a bicycle beat them all for fast, efficient locomotion. Using this analogy, he would point to computers as ways of augmenting human capability and intelligence.
If the home computers of the ‘80s were bicycles for the mind, modern AI systems are jet aircraft. They allow human capability to soar to new heights, and extend well beyond its original capability.
The most powerful combinations of all are the right mix of human and AI. Much has been made of how AI will necessarily replace human workers. Even a moment’s thought reveals the weaknesses in such viewpoints – after all, any time mechanisation or automation has been introduced into an industry, it has been applied to the most dangerous or repetitive forms of drudgery, freeing up resources for more creative pursuits that create more value.
So far, the human resources impact of the latest wave of digital transformation has been largely confined to two broad categories: first, ‘front of house’ client or customer engagement roles in sectors impacted by the shift to online interactions; and secondly, business process roles that primarily involve repetitive tasks, classification, data entry or checking.
In that first category, we can immediately recognise that the rise in online shopping has reduced the need for retail stores, and therefore retail staff. The increased use of chatbots and other machine-learning powered smart interfaces extends that to some customer support or retail assistance roles. The authors identified this trend some while ago in our article “Unexpected Human in the Bagging Area: The Impact of Automation on Retail Workforces”, and considered how this might lead to a rise in ‘experiential’ retail strategies and settings.
In the second category, sectors such as financial services and insurance that have long relied upon credit checking, validation, fraud detection and other pattern matching / data checking have embraced machine learning systems. As systems become more sophisticated, the ability of these systems to do more, and with relatively less oversight or intervention for errors or difficult-to-classify cases has seen a rise in productivity coupled with a reduction in personnel required for the more repetitive tasks.
These impacts are both the result of transformations delivered by machine learning systems running on centralised servers. In the first case, the online shopping or support involves a remote customer interacting over the internet with central servers. In the second, the back-office relevant functions are processed using AI on centralised servers.
This is a local AI, for local people…
Increasingly ubiquitous high-speed, low-latency connectivity means that this client-server model of AI delivery has become commonplace across many fields. We’re used to relying on everything from smart recommendation engines for our media consumption to smart tagging and face/object recognition in the photo library on our phone or computer, and (of course) voice interactions with an increasing array of voice assistants. For the most part, these services involve data being uploaded to central servers, processed and the result being relayed back to our local device. Where connectivity is good, this works seamlessly, but where connectivity isn’t good, the service often fails completely.
We’ve probably all had the experience of suddenly not being able to access an online mapping service when off the beaten track because our phone didn’t have signal. In the genteel setting of the countryside in the UK this tends to be merely annoying, in a true wilderness it can be a matter of life and death. That’s the reason that most vehicle satellite navigation systems have maps built in – local data stores rather than relying on phone signals that might prove illusive when mapping and navigation is most needed.
The same factors have tended to restrict AI deployments in remote locations – uncertain connectivity would rob the systems of utility if they were cut off from server-side processing. The march of progress, and the not-quite-still-holding-true doubling of transistors predicted by Moore’s Law, means that capability previously requiring a dataroom can now be delivered on-device. New silicon designs including highly parallel processing components targeted at accelerating AI, combined with multi-core systems featuring both low- and high-power cores are taking AI processing from the centre to the edge. This new ‘edge device’ capability opens up a radical world of new possibilities.
Especially in industries like agriculture and mining, that have tended to take place in areas with poor connectivity, highly capable edge devices with local processing are bringing the benefits of digital transformation. In other industries – healthcare, logistics, infrastructure – where connectivity might not normally be a problem, but even fractional downtime can be, edge devices are also playing a much-needed role.
These AI deployments are bringing the benefits of that ‘jet engine of the mind’ capability to new places and new categories of worker. With that change come new challenges for managers and HR teams.
Changing roles, changing expectations
The introduction of edge computing into new sectors will bring a number of changes for both employees and HR teams. HR teams will need to prepare for a reduced need for certain roles whilst new roles are identified that require different skill-sets. Planning for such changes will require HR teams to ensure appropriate training is provided and that communication plans are effective.
Frequent headlines in the media warning that AI will lead to huge job losses mean employees may be wary and concerned when their company announces it is implementing a new AI system. Therefore HR teams will need to ensure that good and loyal employees do not decide they would be better off walking before being pushed when the company wants to retain them, albeit perhaps in a different role. The more that companies can give existing employees new skills through training the better for the company, as this will reduce redundancy costs and retain loyal and talented employees.
As an illustrative example, we have seen a leading car manufacturer streamline its build process by using robots and cobots and then focus on developing its employees to carry out the customisation of high-end vehicles. They believe employees are better at performing the customisation process than AI systems and the product personalisation industry is forecast to be worth $9.1 trillion by 2030. This has lead to a shift in focus in their recruitment and training. It no longer needs as many employees on the production line but has identified a need to develop a highly specialised sales team to meet customer demand and maximise profit.
The traditional structure of a company’s work-force and employee skill-sets will change in the industries that are affected by edge computing. The training needs of the work-force will alter as jobs evolve due to use of AI and similarly the skills that the company is looking for at the recruitment stage will change. Character, including adaptability, will become a key focus as employees learn to harness the benefits of the AI systems and their skills and experiences change. McKinsey predicts that by 2030 around 14% of the global workforce will have to retrain.
The advantages that edge devices will bring to industries not so far served by AI will be widespread. We cited agriculture as one such sector above. Drones are able to detect when a crop is ready to be harvested (or indeed planted) as opposed to a farm worker having to regularly travel to check the field, saving man-hours and ensuring the farm harvests the crop at the optimum time. Farms adopting this technology will be better placed to sell their crop as a premium product and use the data collected by the drone and processed by a local edge device to evidence the quality of the harvest.
The use of AI systems and edge processing will inevitably mean large amounts of data are collected – both in relation to the product and the work-force. The augmentation of a work-force with AI systems will improve efficiency in several ways. Whilst it can perform certain roles (the drone monitoring the ripeness of crops), it can enable employees to carry out other tasks more efficiently (Microsoft’s virtual personal assistant being one such example). Businesses will therefore have much more data about the efficiency and output of their employees.
The employees working daily with these systems will need to be made aware of how much data the systems collate and store. The data collation will go beyond simple measures like productivity but can extend to analysing email/message content to identify disgruntled employees or those who are seeking to damage a business. Many employees will have an instinctive fear of such monitoring and how their data is used – not least given the regular incidents of data breaches. Therefore HR teams will need to consider very carefully how such data is stored and accessed to address concerns that employees are being continually monitored by the AI system.
Security of data will be vital and proper procedures for accessing and evaluating the data will need to be followed to allay concerns that the data is used by managers for improper purposes. In addition to ensuring adequate policies are in place, HR teams may need to consult with their employees to retain trust and deal with challenges from the work-force.
A recent Goldman Sachs report stated that AI augmented companies enjoy 28% better performance than their competitors. Businesses will find that using AI systems mean they are able to produce more with their existing work-force. Taking a metaphor from the agricultural sector, the upfront cost of the AI system will bear fruit.
Holding on to Augmented Talent
Employees working with AI systems will see their productivity and profitability increase. Specialised skill-sets and experience of working effectively with AI systems will increase those employees’ value to a business. Deciding the size of pay increase to give to such employees will not be the only point for HR teams to consider. They will also need to ensure the business retains key employees – will salary be enough or will employers have to provide more to retain focused key workers – working flexibly, both in terms of location and hours, training, career development, business values and incentive arrangements will all need to be considered.
Employee management will be key to retaining a highly productive and skilled work-force in a business powered by AI. Good management will no longer suffice, instead HR teams will need to ensure great managers are in place for high performing teams turbo-charged by augmentation from AI. Leadership will need to be emotionally intelligent as reduced face-to-face contact remains normal. Authenticity will be a hallmark of effective leaders. Businesses will need to consider whether they have valid restrictive covenants with key employees to protect their legitimate business interests; that their intellectual property rights and confidential information are protected; and challenge themselves to remain/become a genuine employer of choice in their field.
The potential for AI systems to help businesses win market-share, better focus their marketing and sales budgets and improve both efficiency and profitability is clear. The effect of these changes on employees and the workforce structure will be profound. HR teams will need to effectively focus on training, recruitment/career developments, working arrangements, talent retention and protecting the businesses become vital to ensure businesses continue to leverage the benefits of AI systems and edge processing. Otherwise, they risk being pushed to the edge by competitors who truly realise that their employees remain the key to success.
Living life on the edge
We’ll be discussing the implications of this exciting time in hardware and the impact of increasingly powerful edge devices and local AI at the European Technology Summit in our ‘Hardware Renaissance’ panel. To find out more and register to attend the summit, visit the event website.
You can find more views from the DLA Piper team on the topics of hardware, AI, systems integration and the related legal issues on our blog Technology’s Legal Edge.
If you’d like to discuss any of the issues discussed in this article, get in touch with Gareth Stokes, Robert Forsyth or your usual DLA Piper contact.