Generative AI - The past, present and future.

Generative AI has rapidly transformed from an emerging technology into a disruptive force. While one moment, it’s hailed as the next frontier of innovation, the next, it’s confidently generating hallucinations that sound plausible but are utterly wrong. It’s an exciting, chaotic, and unpredictable space, so let’s take a journey to explore where we’ve been, where we are, and what the future holds for Generative AI.

Watch a full recording of this article, presented by our Lead Consultant, Josh Smith at the NZ AI Roadshow.

How Did We Get Here?

To understand how generative AI exploded into prominence, we need to look back at four major technological eras:

1. Compute Era (1950s–1990s): Laying the foundation for modern technology.
2. Internet Era (1990s–2000s): Connecting the world and information.
3. Cloud, Social & Mobile Era (2000s–2020s): Revolutionising how we interact with data.
4. Generative AI Era (2020s+): Redefining human-machine collaboration.

In this new era, sectors like finance, IT, government, and retail are all leveraging AI to boost efficiency, automate workflows, and enhance customer experiences.

The Last Two Years: An AI Arms Race

In November 2022, ChatGPT made its debut, quietly launching to the public and rapidly becoming the fastest-growing user base in history, amassing over 1 million users in less than five days. Today, ChatGPT has over 180 million users, signalling a seismic shift in how we interact with AI.

By March 2023, the competition was in full swing. Google’s Bard entered the scene, followed by Anthropic’s Claude. Suddenly, an AI arms race was triggered, with companies scrambling to gain an edge. Nvidia, traditionally a GPU manufacturer, rebranded itself as an AI accelerator, briefly overtaking Microsoft as the world’s most valuable company.

The year wasn’t without drama. In November 2023, Sam Altman, the CEO of OpenAI, was fired amidst claims that he wasn’t consistently candid in his communications. In a surprising twist, Microsoft’s Satya Nadella hired him, only for OpenAI’s employees to threaten mass resignation. Ultimately, Altman returned to OpenAI, illustrating the high-stakes leadership crises within the AI industry.

Around this time, AI wearables also emerged, marketed as the next generation of digital assistants. However, these devices, like the Rabbit R1, were met with harsh criticism, with reviewers like Marques Brownlee calling them “barely reviewable.”

As the space matured, major players like Amazon, Microsoft, and Google introduced services like Amazon Q for Business, Microsoft Co-Pilot, and Gemini. Meanwhile, regulators began to catch up. The EU passed the world’s first binding AI regulation, while in California, a contentious AI bill (SB-1047) was introduced.

The Lighter Side of AI: Hiccups Along the Way

It hasn’t all been smooth sailing for AI. Missteps, sometimes comical, illustrate the growing pains of early adoption:

• A Chevrolet dealership chatbot went rogue, offering a car for just $1 in late 2023

• Google’s AI suggested using non-toxic glue as a pizza topping, pulling data from an 11-year-old Reddit post.

• Air Canada’s chatbot gave a customer incorrect information about bereavement fares, sparking a legal battle in which the airline claimed the bot was a “separate legal entity.”

• Pak’n’Save’s Meal-Bot advised customers to make chlorine gas by mixing bleach and ammonia—a recipe that made headlines for its dangerous consequences.

Advancements in Gen AI: A Changing Landscape

Despite these hiccups, there have been tremendous strides in Gen AI, fundamentally reshaping industries. AI is now bridging the gap between robots and humans, aiding in training simulations, and interacting more intelligently with the real world.

Nvidia’s GR00T project is an AI foundation model designed to help robots adapt to their environments, while multimodal models like GPT-4 and Claude 3.5 allow AI to process not just text, but images, sounds, and videos. Tools like Sora are pushing the boundaries with text-to-video generation, and Meta’s AI Studio is letting content creators craft AI-driven digital twins to engage with their audience.

Businesses are keen to integrate AI, reminiscent of the early days of cloud adoption. Startups and enterprises alike are now rushing to tick the “AI” box, whether they’re ready or not.

The Present: Where’s the Value in Generative AI?

Today, generative AI presents immense opportunities in several areas:

1. Content Creation: AI tools are being used to generate articles, images, audio, and video. Platforms like Chat-GPT, DALL-E, ElevenLabs, and Synthesia are reshaping creative industries, making content production faster and more accessible.
2. Customer Experience: Businesses are leveraging AI to improve customer interactions. For example, One NZ eliminated 1 million contact centre calls by using Amazon Connect to analyse sentiment and automate responses.
3. Insights & Knowledge Retrieval: Gen AI has revolutionised knowledge management. Techniques like Retrieval Augmented Generation (RAG) allow organisations to infuse LLMs with their own data, enabling advanced semantic search capabilities that go beyond rigid keyword matches.
4. AI Agents: Autonomous AI agents are transforming workflows, from product design to event planning. These agents can manage multiple tasks—research, analysis, and even decision-making—while improving productivity.

Future Horizons: What’s Next?

As generative AI continues to evolve, what does the future hold?

• Domestic Robots: Elon Musk predicts that within 7 years, most households will have a domestic robot.
• Shift in User Behaviour: People will grow accustomed to assigning tasks to AI agents and waiting for responses, even if it takes days or months.
• Custom LLMs for Business: As AI costs drop and performance improves, custom AI models will become more accessible, transforming sales, customer interactions, and training.
• Redistribution of Jobs: AI won’t take our jobs but will enhance them by automating repetitive tasks.
• Personalised Everything: From entertainment to education, AI will tailor experiences to individual needs, providing world-class education for all.

However, challenges remain. How can we trust AI-generated media? How can we secure intellectual property in an AI-driven world? And will regulation strike the right balance between innovation and protection?

Generative AI is not magic—it’s a powerful tool for human enhancement. As we navigate this new era, the possibilities seem endless, but so do the questions we need to answer. The journey is just beginning.