Articles

Impact of Generative AI -
Trends, technologies & outlook

- Panchalee Thakur

Generative Artificial Intelligence (Gen AI) has taken businesses to the edge of a new frontier that promises unprecedented levels of productivity, innovation, and new revenue streams. With advanced abilities in powering human-AI interactions, content generation, software development and problem-solving, Gen AI is a technological breakthrough that is reshaping business models and driving use cases across the industry spectrum.

Gen AI was ranked first in Gartner’s list of Top Strategic Technology Trends in 2022. McKinsey estimates the economic potential of Gen AI to be in the range of USD 2.6-4.4 trillion by 2040.

The interest in Gen AI is understandable - while the primary role of traditional AI is data classification and pattern detection, Gen AI creates new content that mimics what a human being would create, but with no human intervention. Traditional AI follows predefined rules and patterns to perform specific tasks, whereas Gen AI uses Machine Learning and Deep Learning (ML/DL) technologies to recognize underlying patterns in the training data and in creating new data. Large Language Models (LLM) that power Gen AI are trained on massive amounts of data and take prompts from users to generate new software applications, text or videos. For example, with just a text or image prompt, a Gen AI tool such as Meta’s Make-A-Video can produce a high-quality video.

With Gen AI poised to revolutionize businesses, let’s look at the top trends and technologies driving its adoption.

Top Gen AI trends

In 2022, Gen AI exploded into the mainstream with text-to-image tools such as MidJourney and Dall-E 2. Bolstered by this accessibility, enterprises are leveraging Gen AI for a wide range of functions from software development to sales and marketing, customer service, and research and development (R&D).

Code generation

Gen AI tools such as OpenAI Codex and GitHub Copilot can help accelerate and automate code creation by suggesting codes for software development and reducing manual coding efforts.

App builders

With Google Cloud’s Gen AI App Builder, developers with limited machine learning skills can now create enterprise-grade Gen AI applications. Gen App builder allows real-time collaboration and integrates with other Google Cloud services.

Cobots

Equipped with Gen AI, cobots or collaborative robots such as the Natural Language Processing (NLP) tool ChatGPT can aid human workers at the workplace by offering relevant and quick responses to complex questions. Contact centers can use ChatGPT-powered cobots to deliver personalized responses to customers in real-time, thus improving the customer experience.

Business process re-engineering

Organizations can leverage hyper automation tools such as UiPath, Jiffy.ai and Soroco that combine the power of Robotic Process Automation (RPA) with AI and ML to identify and automate complex business and IT processes.

Synthetic data generation

AI-generated artificial or synthetic data tools such as synthesized.io and Gretel AI mimic real-world data that can be used for ML model testing while preserving data privacy.

Key technologies and models

Inspired by the human brain, Gen AI is a subset of DL technologies that use neural networks with artificial neurons that can understand and manipulate raw data to generate new complex and realistic data. Businesses can use different Gen AI models to meet their specific purposes.

Variational autoencoders (VAE) model

The VAE generative AI model first compresses data into a compact form and then picks the highlights of this data to generate new data. Use cases for the VAE model include anomaly detection, image generation and data compression.

Generative adversarial networks (GAN) model

The GAN model has a generator that creates realistic data and a discriminator that helps distinguish real data from realistic data. Through this process, the GAN model can learn and produce convincing artificial data such as photorealistic images for creative purposes.

Auto-regressive model

This Gen AI model generates data step by step, predicting each part based on the data created in the previous steps. The auto-regressive Gen AI model can carry out tasks such as text generation and music composition.

Flow-based model

Converting simple data distributions into complex ones to generate realistic content, this Gen AI model helps with tasks such as image generation and density estimation.

Transformer-based model

Used typically for content creation, conversation and language translation, the transformer-based model uses DL architecture to read and generate human-like text.

Gen AI adoption challenges

While Gen AI has ushered in a new era of business innovation, a recent IDC survey shows that only 30% of enterprises are ready to adopt it. Survey findings indicate that two out of three top business leaders are concerned about data bias and disinformation associated with the use of Gen AI.

Limited dataset availability is another challenge limiting the adoption of Gen AI. A case in point is that of the healthcare industry, where creating datasets comprising diverse patient information and clinical notes can be cumbersome and time-consuming.

Gen AI also has poor interpretability, which makes it difficult to understand how the technology generates outputs. For example, regulators and auditors in the financial industry may need an explanation of financial forecasts made by Gen AI models before accepting the results.

Data privacy and accuracy are other key concerns hindering Gen AI adoption. However, collaborating with data creators and attributing content can help business leaders address copyright infringement concerns. Moreover, businesses can secure their data against leaks via encryption and tackle deepfakes by classifying and protecting sensitive data.

The adoption of Gen AI has also given rise to concerns about its impact on human resources and the environment. To counter job risks, businesses must invest in employee reskilling to ensure they stay relevant.

Exploring Gen AI

Adopting Generative AI (GenAI) can be a game-changer for enterprises. By leveraging the technology, businesses can streamline processes, enhance creativity, and provide more personalized experiences. However, successful GenAI adoption requires careful planning, collaboration, ethical considerations, and a focus on data quality. Starting with small pilot projects, investing in employee training, and ensuring security and compliance are vital to ensure a strong return on investment.

For successful Gen AI implementation, companies will need to evaluate their data architecture, technical know-how, technology infrastructure, risk management and business models. By taking care of these aspects an organization can create the right environment for the ethical and responsible use of Gen AI.

Request a consultation
About the author

Panchalee Thakur

Independent Consultant