The Impact of AI on the UK Workforce
Summary Notes of Event
Event Description:
Hosted by the Future of Work
01/04/19 @ the University of Bath School of Management
Speakers:
-
Prof. Veronica Hope-Hailey, Dean of University of Bath’s School of Management
-
Ravin Jesuthasan, author of ‘Reinventing Jobs’ (2018)
-
Juani Swart, University of Bath
-
Dionysios Kola, Barclays Bank
-
Cihan Kurt, GSK
-
Vaggelis Giannikas, University of Bath
Prof. Veronica Hope-Hailey - Introduction:
Example of the need to future-proof the school buildings:
-
There will be changes in education, studying, and even existence in the coming decades.
-
Nature of change from tech and AI will be fast; and will replicate or replace human skills
-
Need to think of replacement of tasks, not jobs -> people and relationships are the preserve of humans, regardless.
-
New jobs being created, e.g.:
-
IT home consultant
-
Auditors of AI
-
Drone operations, e.g. centralised lorry drivers
-
-
Challenges:
-
Managing the human transition
-
What happens to those left behind?
-
Brexit given as example of what happens when swathes of people don’t enjoy the benefits of globalisation and lash out against the system.
-
-
Manage the disruption of 5-10 years of experimentation
-
Disruption can be costly to business as they adapt.
-
-
Surveillance and security issues?
-
Implication for government regulation on AI’s role in society
-
Ethical considerations.
-
-
-
Answers?
-
Values and ethics
-
Wisdom and judgement
-
Education, learning and research
-
Ravin Jesuthasan - Optimising Automation and Human Labour:
Skills as the currency of the job market
-
‘Legacy’ was an asset and is now a liability ahead of the 4thindustrial revolution.
-
Tech growth has come about due to:
-
Enabling infrastructure
-
Cloud-based storage
-
large amounts of data
-
processing power
-
-
power of convergence
-
speed of adoption
-
-
Industrial Revolutions:
-
2nd– in the 19th-20th century
-
Assembly line model
-
Organise work into jobs
-
Jobs as careers
-
Amplification of labour
-
E.g. General Motors
-
Great cost, great certainty
-
-
3rd– 60s-90s
-
Nikefication
-
Democratisation of information
-
Tech enablement and use of web
-
Enables outsourcing
-
Companies as nexus of contracts
-
-
4th– 2000s
-
Uberisation
-
Mobiles, sensors, machine learning
-
Companies as platforms
-
Work as activities
-
Talent on demand
-
-
2 scenarios:
-
Human work becomes obsolete <-> Human work is constantly reinvented
-
Need UBI; Humans as consumers <-> Gigs, projects, tasks; workers evolving; reskilling as key to economic status and progress; access to reskilling essential
-
Employment relationship changing
-
Moving from selling products to selling subscriptions.
-
Jobs that were outsourced now being brought back in (e.g. customer service/call centre)
-
Regulation was fit for 3rd industrial revolution -> needs to be overhauled.
-
Regulation as reaction
-
Doesn’t look ahead.
-
-
E.g. China 15% of workers are gig workers. Wild West of regulation that is slowly changing and government is reacting by breaking up market concentration
-
Uber ‘flexible fund’ on top of cost for labour/gig to be used for e.g. pension or re-skilling.
-
Washington state rolling this out, extending also to New York.
-
Negotiated from a position of power rather than as a reaction.
-
-
Idea to give Facebook, and platforms like it, a license for 20 years of operation before it gets democratised.
-
It’s then up to them to stay relevant after
-
-
Automation is the tail that wags the dog of the organisation.
-
Example of staff laid off because system that was developed that automated case claims -> ultimately lead to loss of customers and ex-employees were hired back as contractors
-
-
Automation will bring new combinations of work, talent, skill requirements, and work relationships.
-
E.g. no-code AI: don’t require python developer
-
Enables non-data scientists/developers to make use of machine learning.
-
-
-
Over the next 3 years:
-
Organisations redesigning roles for those with more skills (45% of organisations) and lower skills (42%)
-
Free agent work to rise by 50%
-
Full time employees set to reduce from 83% to 77% globally.
-
-
Skills passports will be used to track the skills gained over time.
-
Universities may increasingly offer re-skilling opportunities for life.
-
There’s a need for middle managers to orchestrate and coordinate between multiple facets of an organisation:
-
Agency
-
RPA
-
AI
-
Volunteers
-
Platforms
-
-
Need to deconstruct jobs and see where automation fits in. Enablers of automation include:
-
RPA (Robotic Process Automation) -> routine high volume
-
Cognitive automation -> non-routine creative
-
Social Robotics -> routine collaborative
-
-
Automation as augmentation.
-
Deconstruct and thenautomate.
-
Substitute, or
-
Augment, or
-
Transform
-
-
In terms of return on improved performance:
-
Failing the minimal standards have consequences that can be damaging to the whole organisation
-
E.g pilot.
-
-
Being a superstar, innovator and creative makes you more productive than being ‘good’
-
E.g. creative programmer.
-
-
-
Case study of oil and gas.
-
Replace task of ‘motor hand’ which involves laying pipes on an oil rig
-
CEO has aim of 100% retention of workers
-
-
Solution is to re-skill motor hand worker as a rig technician:
-
Set up, calibrate, and maintain the machine.
-
-
However, people might be resistant to reskilling; unwilling to be schooled/educated.
-
Perhaps lower skilled jobs taken by those who were originally unhappy with or hostile to the education system.
-
-
-
Activities will:
-
Be centralised
-
Shift to other roles
-
Be augmented
-
Be eliminated
-
Replaced by new ones
-
-
Emerging pivotal skills:
-
Orchestration
-
Curation
-
Enabling
-
Juani Swart - A knowledge-based perspective:
How our jobs evolve with AI/automation.
-
There is data that jobs will be lost and data that jobs will be replaced.
-
Social care and education faces labour shortages
-
AI will impact on jobs, not replace.
-
Need to re-draw knowledge boundaries
-
“what do we do?”
-
“how do we do it?”
-
​
-
Impact:
-
On routine jobs
-
High
-
-
On knowledge-based jobs
-
Augment
-
-
Social jobs
-
Low to no impact -> future of work is here.
-
-
-
2 scenarios:
-
1: Everything stays roughly the same but with new tech that is disruptive and augments work.
-
2: Shift to experience economy, much more disruptive effect of automation/AI.
-
E.g. health, beauty, travel, care, etc.
-
-
Dionysios Kola:
Definitions:
-
Process understood as activity that transforms an input to an output.
-
Automation as deploying the digital workforce to do time-consuming and mundane tasks; physical workforce to focus on value added activities.
-
RPA: configurable software that does the task assigned to it by the user; automate by imitating a humans interaction with an application.
-
Workflow automation: enables discovery, modelling, rapid change, governance and end-to-end visibility of processes; process visualisation and basic team management, point to point management.
-
AI/Machine Learning: uses experience to achieve expertise in a domain.
-
Doing <-----> Thinking
-
Rule-based <-> jugement-based
-
Robotics <-> ML/AI
-
-
Benefits of implementing automation
-
Data analysis and insights
-
Reliability
-
E.g. can work 24/7,
-
-
Cost benefit savings
-
Reduction of operational costs
-
Cost avoidance
-
Increase in sales
-
Shouldn’t replace all workers as this leads to being less competitive
-
-
Increased productivity
-
Non-invasive tech
-
No disruption to underlying legacy systems.
-
E.g. banks.
-
-
-
Accuracy
-
Less prone to manual errors
-
-
Increased employee satisfaction
-
Less menial, repetitive tasks
-
-
-
Uses (case study of banks)
-
Price-scanning and setting of e.g. bond prices, faster.
-
Contract comparison: check if contract is standard, freeing up lawyers
-
Background checks and risk assessment
-
Monitor and prevent criminal and fraudulent activities
-
Chatbots
-
Key take away is to stay relevant and be resilient.
Cihan Kurt - AI and Customer Service:
(1) understand your customer
-
move from return on investment -> return on experience
-
we’re living in an age of personalisation
-
Need to consult what the data tells us.
-
Data includes:
-
Demographics
-
Transactions
-
Market research
-
Social media engagement
-
-
Descriptive (what’s happening?) àDiagnostic (why is it happening?) àPredictive (what’ll happen next?) àPrescriptive (what do I need to do?)
(2) AI in interactions
-
Used to target; predict; recommend; personalise.
-
E.g. Natural Language Generator
-
Turns a set of data into natural language text that summarises the data in an easy to read way.
-
-
Health chatbot
-
(3) AI in sales
-
Used for analysis; to deconstruct sales channels; marketing
-
Optimise strategy
-
Redistribute resources
-
Calculate return on investment
-
Sales forecasting
-
Dynamic real-time pricing
-
Next best action suggestions
-
Resource optimisation
-
Vaggelis Giannikas -
AI in Industrial Operations:
-
AI can help in analysis and decisions– sits there mostly
-
Lots of people work in sense and act
-
Tech can help in this domain, AI less so
-
AI understood as decision-making software
-
-
-
A lot of Industrial Ops is not decision-making
-
AI is not automation, though they do go together.
-
It’s worth understanding what digital things can and cannot do.