Generative AI – Delving Beneath the Effervescent Surface

Generative AI – Delving Beneath the Effervescent Surface

The disruptive potential of generative AI and its speculated impact on the job market have been widely discussed. However, the intricacies of generative AI remain complex and often confusing. It is crucial to gain a comprehensive understanding of this topic and not develop colored opinions. In this first of a multi-part series, we aim to provide a concise overview of generative AI and delve into its key aspects to help develop an informed perspective.

What is Generative AI (GenAI)?

Gen-AI refers to a category of artificial intelligence technologies capable of producing various forms of content, including text, images, audio, and synthetic data (artificially manufactured data sets for use in testing of models). It must not be mistaken for automation or general AI. Automation involves following pre-programmed rules, while general AI aims to simulate human-like thinking. Generative AI, a subset of AI, pertains to intelligent machines capable of producing novel content.

A Brief Evolutionary Timeline of Generative AI

  • The Markov Chain, an early example of Generative AI, employed statistical models to generate new data sequences based on input.
  • The creative potential of Generative AI is epitomized by Generative Adversarial Networks (GANs), a specific type of neural network proposed by Ian Goodfellow and colleagues in 2014.
  • WaveNet (2016) introduced realistic human speech generation, leading to more human-like AI assistants and highly accurate text-to-speech
  • OpenAI’s Generative Pre-Trained Transformer (GPT) models (since 2018) marked a significant advancement in text-based Generative AI.

Among several industry verticals, GenAI finds wide applications in the software industry for code generation, documentation, and quality assurance, as well as in creative fields like advertising, enabling efficient content creation for blogs and social media. In the pharmaceutical and medical sectors, GenAI aids in drug discovery, design, diagnostics, and imaging. Moreover, it plays a role in finance, assisting in fraud detection and risk management.

So, Where Do the Concerns Lie?

While GenAI helps with repetitive tasks, apprehensions arise about potential job displacement. The fear is that it might jeopardize people’s livelihoods. McKinsey’s Jun 2023 research suggests that GenAI could automate 50% of global activities from 2030 to 2060. Goldman Sachs predicts AI could replace around 300 million full-time jobs. However, the report also acknowledges that 60% of current occupations didn’t exist in 1940. Thus, the situation might not be as dire as predicted. Nonetheless, historical trends suggest that technological advancements can lead to job losses in the short term.

Recent layoffs attributed to GenAI include the May 2023 report by Challenger, Gray and Christmas, which indicated 80,089 job cuts in the USA, with about 4.9% (3,900) directly tied to AI. A survey by revealed that almost half of surveyed US companies had adopted ChatGPT, leading to worker replacement across various functions.

Real-world Impact

An individual’s story, as reported by The Washington Post, sheds light on personal impact. Eric Fein, a content writer, saw his contracts canceled as companies adopted ChatGPT for content creation. This reflects a broader concern about job security.

However, the generic assertion that Generative AI will unequivocally eliminate jobs is flawed. The Aug 2023 International Labor Organization study suggests that most jobs and industries will be supported rather than substituted by automation. Only a fraction of employment faces substantial automation risk (5.5% in high-income countries and 0.4% in low-income countries), and the focus should shift to potential changes in job quality, intensity, and autonomy. Torsten Bell, CEO of the Resolution Foundation, cautions against placing undue certainty on AI’s long-term impact, advocating for a measured perspective.

However, industries dependent on “programming and writing” skills are certainly the most vulnerable to GenAI advancement. Research from both Forbes and Goldman Sachs predict that jobs in legal services are likely to be highly impacted due to GenAI as it is expected to make services cheaper and accessible to the masses as well as play an important role in referencing historical court cases for latest judgements. Banking & Financial operations are already witnessing increased influence of AI. Morgan Stanley has started using AI-powered chatbots to organize its wealth management database. The World Economic Forum predicts AI will impact the finance division in that it will promote both job cuts and job creation coupled with increased efficiency. Media, Marketing & Administrative Support services (which include backend BPO operations as well), which involve common tasks such as writing reports, news, content, and data input are very likely to be fully automated. Business Insider, CNET, BuzzFeed and CNBC reporters have all used ChatGPT to write articles in the past and other media houses are expected to follow.

On the contrary, jobs in industries that rely on manual labor are least exposed to GenAI. A Mar 2023 report by Goldman Sachs Global Investment Research suggests Buildings, Grounds Cleaning & Maintenance (Facilities Management), Installation, Maintenance & Repair (Customer Service & Support, After-sales support); and Construction & Extraction (Infrastructure, Mineral & Mining) are industries where jobs replaced by Gen AI will be the least at 1%, 4% and 6%, respectively.

Transformative Potential, Albeit with Apprehensions

Generative AI can reshape business processes, enhance productivity, and streamline tasks. Its applications span from personalizing e-commerce experiences through chatbots to optimizing logistics, power usage, and predictive maintenance.

However, considerations regarding GenAI extend beyond economic, financial, and labor concerns:

  • Creativity and critical thinking might suffer as reliance on AI replaces individual contemplation.
  • Generative AI’s output is limited by historical and potentially biased data, lacking the nuanced ethical reasoning of humans.
  • Ethical and copyright concerns arise due to potential misuse for generating false information, raising questions about responsibility and intellectual property rights.

Generative AI is still at a very nascent stage and its impact is uncertain. While it could offer benefits, regulatory and ethical frameworks are crucial. Governments, companies, and users must approach its development with cautious optimism. In its current form, Generative AI, while transformative, cannot replicate human resilience and sentience. Striking a balance between innovation and safeguards will determine its role in shaping our future.

Author: Shivam Agarwal,

Assistant Consultant, Strategy Consulting

Image courtesy: Wikimedia

No Comments

Post A Comment