Headcounts on the block? The AI revolution takes hold
Artificial intelligence (AI) has the potential to revolutionize the way we work and automate many tasks that currently require human labor. Some experts believe that AI will lead to significant job displacement in the coming decades, while others think that it will create new job opportunities and boost productivity.
Sounds convincing, doesn’t it? Even more so when you know that a human being didn’t write that paragraph. AI wrote it. Or more specifically, ChatGPT did.
The precocious chatbot from OpenAI unveiled to great acclaim in November 2022, uses a large language model that harnesses natural language processing (NLP). This is a branch of AI that seeks to analyze, comprehend, and communicate with language.
The GPT-3 (Generative Pretrained Transformer 3) AI was trained on massive amounts of data and user feedback to sound like humans and can generate anything from recipes to resumes on demand (even in the style of Shakespeare, if it pleases thee). Translation, sentiment analysis, introductions to articles about AI… ChatGPT can even create code from simple text prompts. It is one of the largest and most powerful language processing AI models to date, with 175 billion parameters.
Computers, it seems, are increasingly infringing on complex tasks once thought quintessentially human (including deceiving professional backstabbers in the game of Diplomacy).
Meanwhile, Microsoft is investing up to US$ 10 billion in OpenAI, the company behind ChatGPT. Professional writers aren’t looking for alternative employment just yet – but they’re definitely typing just a little faster!
From the printing press to the PC, to machines that read, write and produce fine art, human history is built on technological innovation and automation. Does our future belong to AI?
The business world certainly thinks so.
AI is a smart move
AI adoption more than doubled between 2017 and 2022, according to McKinsey’s State of AI 2022 survey, with 50% of respondents adopting AI in at least one business area. The average number of AI capabilities per organization has also doubled, from 1.9 in 2018 to 3.8 in 2022. Robot process automation and computer vision remain top of the list, while NLP has leaped from the middle of the pack in 2018 to third place.
Why the explosion of interest in AI?
AI can analyze vast amounts of data, automate processes, and provide insights and recommendations that enable businesses to enhance accuracy, efficiency, and productivity, improve decisions making, increase personalization and customization, and streamline people management. The list goes on.
AI can drive revenue, efficiency and cost savings across multiple domains. In 2018, manufacturing and risk were the juiciest targets. Now sales and marketing, product and service development, and strategy and corporate finance are delivering the biggest gains, with the deepest cost-cuttings happening in supply chain management.
According to a PwC report, AI could contribute up to US$ 15.7 trillion to global GDP by 2030. An innovation injection of this magnitude might reverse dwindling labor productivity and revive the global economy, which still bears the scars of the COVID-19 pandemic and the financial crisis of 2007-2009.
Industry use cases include:
- Healthcare: Analyzing medical images, predicting patient outcomes, and improving the accuracy of diagnoses (e.g., analyzing X-rays and identifying abnormalities that may indicate a particular condition). Pharmaceutical companies like Moderna used AI to develop COVID-19 vaccines, and in 2020, the Jameel Clinic at MIT identified Halicin through machine learning and AI models.
- Finance: Analyzing financial data, making investment recommendations, detecting suspicious transaction patterns, and preventing fraudulent activity. Business Insider estimates that AI could save banks and corporate institutions US$ 447 billion by 2023.
- Manufacturing: Improving quality control, optimizing production, and predicting maintenance needs (e.g., monitoring production equipment and identifying patterns that could indicate potential failures). Tesla’s Gigafactory is a marvel of AI-driven automation.
- Retail: Personalizing customer experiences, optimizing pricing and inventory management, and improving marketing (e.g., analyzing customer purchase data and recommending products to individual customers based on their past purchases). Amazon Go’s white label “Just Walk Out” technology is enabling multiple retailers to experiment with cashier-less stores.
- Transport and logistics: Managing supply chains, predicting demand, instructing warehouse robots, optimizing routes, improving safety, and reducing fuel consumption (e.g., analyzing real-time traffic data and recommending the most efficient journeys for delivery trucks). Walmart used AI supply chain simulations to rapidly reroute deliveries and predict demand changes after Hurricane Ian forced the closure of a major distribution hub in Florida.
- Sales and marketing. Creating content, generating leads, improving customer experiences, and managing customer relationships (e.g., analyzing data to improve targeting and delivering greater personalization via chatbots). Salesforce’s Einstein Analytics applies AI to anticipate customer behavior and make recommendations.
Abdul Latif Jameel is already exploring how to leverage the benefits of AI for the communities and markets it serves across the globe. AI is at the heart of a number of healthcare technologies that Abdul Latif Jameel Health has invested in, for example. These include a handheld ultrasound device developed by Butterfly iQ+TM and extended reality (XR) 3D virtual environment surgical technology developed by Holoeyes.
Another Abdul Latif Jameel business, Fotowatio Renewable Ventures (FRV) (part of Abdul Latif Jameel Energy), through its innovation arm FRV-X, also harnesses the power of AI in deploying the Tesla auto-bidding software used in its battery storage systems (BESS) in the UK (Holes Bay, Dorset; Contego, West Sussex; Clay Tye, Essex) and in Australia at Terang, Victoria, and a hybrid plant at Dalby, Queensland.
Also part of Abdul Latif Jameel Energy, Almar Water Solutions has invested in a technology company that provides AI-driven IoT products and services for digital transformation in the water, energy and mobility sectors: Datakorum. Datakorum subsequently secured a five-year project with leading communications operator, e& Enterprise, (previously Etisalat Digital) facilitating the digital transformation of the water and energy management systems in the City of Abu Dhabi, UAE.
The AI playbook
AI leaders are already pulling ahead, and laggards will find it increasingly difficult to catch up. Why? According to McKinsey, they have a first mover advantage, take the long view, invest more and more wisely, attract the best talent, and follow an emerging set of principles to embed and scale AI. These ‘foundational practices’ include linking AI strategy to business outcomes, understanding how to integrate AI into business processes and decision-making, prioritizing talent acquisition and training, and ensuring AI and people work together to create more value.
There are also a number of ‘frontier practices’ that enable AI development and deployment at scale in the so-called ‘industrialization of AI.’ These include:
- High-quality data, modular data architecture that can flexibly accommodate new applications and automate data-related processes.
- Engaging non-technical employees with low-code or no-code programs.
- Scaling with standardized tools and in-house end-to-end platforms.
- Assetization of code for reuse in different applications.
- Mitigating risk by developing policies for privacy and fairness, for example.
At least latecomers don’t have to write the book from scratch.
Better managers, better work?
AI can monitor hazardous conditions and improve predictive maintenance to create safer working environments (or remove humans from dangerous areas entirely). It can also be used to enhance training and development. For example, tech giant Honeywell uses AI-based image recognition technology to create virtual reality training courses for new hires.
AI also creates jobs, from AI-specific roles to new ones created by AI-driven growth. By automating a wide range of tasks, AI allows traditional carbon-based lifeforms (i.e., humans!) to focus on more complex and creative tasks, including occupations that don’t exist yet. It can even automate aspects of recruitment and hiring, performance evaluation, and employee engagement. For example, using chatbots to interview candidates and analyzing data on attendance, productivity, and customer satisfaction to provide a more objective assessment of performance. So-called ‘algorithm management’ has the potential to streamline management and protect against biases. However, it could also reinforce discrimination and infringe on workers’ rights, as we’ll see.
Brave new world, same old worries
Before the robots wake up and extinguish the human race, as some doomsayers would have us believe, businesses have more pressing hurdles to worry about, most of which are still being worked out.
Firstly, AI is not easy. Algorithms require massive training data and can be difficult to generalize across use cases. And talent is scarce. Businesses need experienced managers and strategists, along with AI-specific and general digital know-how, including software engineers that can develop intuitive front-ends for non-techy users to interact with algorithms.
As for other jobs…
Some will change. Others will disappear. Manual labor in predictable environments is first in the firing line, but even knowledge workers are at risk. For example, AI is already helping radiologists analyze images – and it’s only getting better. What happens when it’s consistently more reliable? And large language models should make anyone who writes for a living a little uneasy.
“We should stop training radiologists now. If you work as a radiologist, you are like Wile E. Coyote in the roadrunner cartoons; you’re already over the edge of the cliff, but you haven’t looked down,” believes Dr Geoffrey Hinton, cognitive psychologist and computer scientist at the University of Toronto and an advisor to Alphabet, the parent company of Google.
Many workers will need to change occupations within organizations and sectors and across geographies. Physical activity in highly structured environments and data collection and processing will decline. Meanwhile, complex activities and jobs in unpredictable physical environments will remain dominated by humans (good news for managers and massage therapists alike).
Advocates argue that AI and humans will work together, as many already do, with humans typically shifting from doers to managers, like when Amazon’s lift stackers becoming robot operators. After all, only 5% of occupations could be fully automated by current technologies.
Others say this is wishful thinking. Technology is improving so quickly, and AI will inevitably replace humans where automation is possible. By some estimates, AI and automation could displace around 400 million workers between 2016 and 2030.
Either way, upskilling and reskilling is essential. Naturally, digital and AI-based expertise are in high demand. But other skills will rise in importance as AI takes over and automates more activities. These include social, emotional, and higher cognitive skills, such as creativity and critical thinking.
Coworkers, rather than competitors?
It’s not all doom and gloom. A Boston Consulting Group (BCG) survey suggests AI is broadly positive for individuals and organizations, with 60% of respondents saying they feel like AI tools are a coworker. In fact, AI is so embedded in business and consumer products it often goes unnoticed. Furthermore, organizations with employees who personally derive value from AI are 5.9 times as likely to get significant financial benefits from AI compared with organizations where employees do not get value from AI. Managers who want to maximize AI ROI should cultivate trust, understanding, agency, and awareness among employees. “AI can improve an individual’s self-determination through greater competency, increased autonomy, and stronger relationships,” says the report.
What about wages?
If AI and automation perform tasks faster, cheaper, safer, and more reliably, demand for human employees could drop, and so too salaries High-wage jobs may increase, but many ‘safe’ jobs that cannot be automated, such as nursing, are lower-paid. This could exacerbate income inequality and political and social unrest.
There will be increased need for financial support and safety nets to offset the impact of automation, such as Universal Basic Income, a stipend for all citizens as championed by former US presidential candidate Andrew Yang. And Governments must create more substantial, sustainable work through infrastructure and climate change adaptation investment, for example.
Ethics vs. algorithms
The use of AI raises several ethical questions and concerns which must be addressed. The following all call for human oversight:
- Equity and fairness. AI systems trained on biased data could lead to unfair treatment of certain groups of people. This can be avoided through diverse and representative training data, robust testing, and accountability measures.
- Transparency and explainability. Many machine learning algorithms, such as foundation models, are black boxes that cannot explain their decisions. “Computer says no” isn’t good enough. US defense agency DARPA is developing algorithms that “explain their rationale, characterize their strengths and weaknesses, and convey an understanding of how they will behave in the future.”
- AI systems must be held accountable for their decisions and actions. Businesses should define the scope and limitations of AI systems and establish clear policies and procedures to ensure they are used responsibly and compliantly, incorporating human oversight and mechanisms for addressing negative consequences.
- AI will deepen workplace surveillance with the ability to analyze written, verbal, and even facial queues to assess workers’ moods and productivity. Businesses must update their data governance policies and procedures to ensure personal data is handled responsibly and in compliance with relevant privacy laws.
- AI and automation generally improve safety by removing workers from hazardous areas. But organizations shouldn’t take any chances, particularly with autonomous vehicles and collaborative robots (cobots). Regular health and safety audits are essential.
- Security and resilience. AI systems, and the data they use and generate, are high-value targets for criminals, intelligence services, and disgruntled employees. And natural disasters can compromise or destroy infrastructure. Cyber-security and resilience are top priorities for businesses and governments alike.
Addressing the risks
Despite widespread AI adoption, there have been no substantial increases in risk mitigation over the last few years. Time for governments and policymakers to step up?
As the OECD asserts in a whitepaper on how to ensure AI remains trustworthy, most concerns can be addressed by applying existing policies and regulations, such as anti-discrimination, data protection (e.g., GDPR), deceptive practices, and rights to due process – and using these as the basis for developing new policies to protect workers.
The EU AI Act seeks to classify and rate AI systems used in employment, banning some and subjecting others to legal requirements over data protection, transparency, human oversight, and robustness. In the US, some states require applicant consent for facial recognition tools in hiring, and the New York City Council mandates algorithmic bias audits for “automated employment decision tools.”
When robots rule the world
AI is a global phenomenon. All countries and sectors stand to benefit, especially those with relatively high wages. By 2030, automation could displace 20% to 25% of the workforce in countries like France, Japan, and the United States – more than double the rate in India.
Owing to AI’s strategic and economic importance, governments in China, America, and the rest of the world are employing various strategies to drive AI adoption. As part of its New Generation Artificial Intelligence Development Plan, China is creating tax breaks and subsidies and increasing R&D expenditure by over 7% each year to become a world leader in AI. The government has also implemented regulations to encourage data sharing among companies to fuel AI development. In the United States, the government is funding academic institutions and research centers and loosening regulations on the use of AI in industries like autonomous vehicles. Likewise, the European AI Strategy seeks to make the EU a world-class hub for AI and ensure that AI is human-centric and trustworthy.
There are also several international organizations working across different levels of government, NGOs, and private sector stakeholders to address AI’s ethical, societal, and economic implications, fund research, and foster responsible development and use of AI. These include the European Union’s Horizon 2020 program, The Global Partnership on Artificial Intelligence (GPAI), The Forum for Cooperation on Artificial Intelligence (FCAI), and the Partnership on AI. The OECD’s AI Policy Observatory tracks and analyzes the AI policies of its member countries, many of which have adopted the OECD’s guidelines for the responsible use of AI. And The UN General Assembly adopted a resolution on the Promotion of the Use of Artificial Intelligence for Development in December 2020.
How should businesses prepare for AI?
Of course, there is no single route to success with AI. Every business will need to find its own route to leveraging the potential of AI to bring new benefits, not just destroy old models.
Perhaps AI itself can provide a solution? This is what ChatGPT had to say when asked how businesses can best prepare themselves for the AI revolution:
All good advice. But no supercomputer can predict the future (yet).
Fortunately, it only takes old-fashioned common sense to realize AI has changed business forever. It’s up to us to ensure those changes make things better, not worse.