Published on
February 26, 2024
AI is changing the world one day at a time while we, humans, are still in the pilot seat. But what if AI takes the pilot seat? Can we foresee it? Can we outsmart it? Where is the line, and are we close to crossing it? We stand at a pivotal moment in the era of AI, where the decisions we make in our businesses and as a society now will have an immense influence over the forthcoming future.
Looking at the state of AI, we can safely say that It has the power to change the course of the world and our lives. As the founder of Pixyle.ai, I find myself at the forefront of this stimulating wave of change. I am excited to share how we are embracing AI, the challenges and promises it holds, and what businesses need to consider as we embark on this transformative journey.
Let’s take a step back and look at AI's timeline:
1950s-1960s: The Birth of AI and the Perceptron
In 1950, Alan Turing introduced the Turing test, suggesting "machines can think." Frank Rosenblatt's 1957 invention of the perceptron paved the way for early artificial neural networks, shaping modern deep-learning models.
The 1980s-1990s: Post-AI Winter and the First Neural Networks
After the "AI winter," the 1980s and 1990s saw a resurgence in AI. The era introduced Lisp machines, expert systems, and the connection machine. Neural networks and behavior-based robotics emerged in the mid-1980s. Marvin Minsky's "The Society of Mind" and milestones like the DART scheduling app during the Gulf War and IBM's Deep Blue winning at chess highlighted the 1990s. Sony's AIBO debuted as an early AI "pet." However, despite Geoffrey Hinton's neural network advancements, limited computational power hindered widespread adoption in this era.
Late 1990s-2000s: Machine Learning Emerges
In the late 1990s and early 2000s, machine learning gained prominence, using statistical methods for data-driven decision-making. Emphasizing practical applications with shallow learning, highlights included interactive robot pets, Cynthia Breazeal's sociable machines, iRobot's Roomba for autonomous vacuuming, NASA's Mars rovers, Honda's humanoid robot ASIMO demonstrating human-like walking, the University of Alberta's success in resolving Checkers, and Google's strides in autonomous cars.
The 2010s: Deep Learning Dominance and the Birth of the Transformer
In the 2010s, deep learning took centre stage, with major strides in image and speech recognition through neural networks. A turning point came in 2012 with Alex Krizhevsky's ImageNet breakthrough, catalyzing the shift from shallow to deep learning. Google introduced the Transformer model in 2017, followed by the Transformer architecture in 2018, leading to influential models like BERT. Google Duplex revolutionized AI-assisted phone appointments, while DeepMind's AlphaStar reached Grandmaster level in StarCraft II in 2019, showcasing AI progress. OpenAI's GPT-2 in 2019 showcased the transformer architecture's power.
The 2020s: Advances in Generative AI
In the 2020s, generative AI has made big strides with influential models like DALL-E, Stable Diffusion, and Midjourney, expanding the creative and adaptive capabilities of artificial intelligence. Also, Microsoft releases Turing Natural Language Generation (T-NLG), a 17-billion-parameter model. AlphaFold 2 by DeepMind wins the CASP competition for protein structure prediction. OpenAI introduced GPT-3 in May 2020, followed by GPT-4 in March 2023, a multimodal model integrated into ChatGPT.
2023: What’s happening now?
A few months ago, on Nov 6th, 2023, OpenAI made significant strides at their Developer Day, introducing an AI marketplace for custom GPTs. Now everyone can build tailor-made GPTs and chatbots for specific tasks and use cases. ChatGPT is now also multimodal, it can see, hear and speak. On another front, Elon Musk’s advancements in AI chips are becoming a reality. Neuralink got approval from the US FDA for its first human tests.
In the fast-moving world of AI, exciting progress has been happening across all industries. We're now in a moment where AI isn't just a sidekick but is evolving toward learning on its own from vast amounts of data, without constant supervision and the need for labeled data. This change is really important when we look at where AI is headed and how it could shake up different industries.
AI in E-commerce
AI is transforming e-commerce. It’s making the online shopping experience engaging and tailored to each customer's needs and preferences. On the other hand, AI is also helping e-commerce businesses and teams be more efficient in their work. Pixyle's AI, for example, simplifies the process of adding products to online stores by efficiently creating detailed tags, titles and descriptions from images. Adobe Firefly makes product image editing super easy for photo production teams. Google Shopping uses Generative AI to easily add different backgrounds to product images. Zalando released AI Style Agents for consumers to talk to and suggest items based on individual tastes and asks. Algolia's vector search improves search accuracy by considering the meaning behind words, providing more relevant results. Together, these companies showcase how AI is revolutionizing e-commerce.
But what’s next? I am confident that, at one point, AI will not only help with product data creation and image moderation but will also be generating complete e-commerce stores and product pages on its own. Something like this is already happening in the web development world. Builder.ai for example, uses generative AI to build apps from scratch in mere minutes.
AI in Arts and Marketing
When it comes to the industry where people are most afraid that AI will replace them, it is probably the arts, creativity and marketing niche. The idea of AI quickly coming up with and running entire marketing campaigns in just 15 minutes is unsettling for marketers. However, Coca-Cola's "Masterpiece" and Nutella's "Nutella Unica" campaigns show a collaborative approach where AI can actually create positive changes, providing more and better opportunities for creative jobs.
AI in Entertainment
Spotify's breakthrough in automatically translating languages in podcasts using voice hints at the broader use of AI in instant voice translations. This innovation could impact traditional translation jobs, as AI accomplishes tasks in seconds that currently take humans a day. Meanwhile, Meta's Emu Video and Emu Edit aim to enhance creative expression and communication, turning text into videos and enabling precise image editing. Generative AI is also transforming Hollywood, seen in the film "Everything Everywhere All at Once," powered by Runway AI. However, challenges like unauthorized data use require attention for responsible AI use, striking a balance between innovation and ethical standards in the dynamic landscape of generative AI.
AI as Your New Employees to Hire
How about this? Artisan introduces fully AI digital workers, employees like Ava and Noah, who seamlessly integrate with human teams. These AI workers automate tasks, optimize workflows, and self-improve daily. With an intuitive dashboard for easy management and the ability to chat for tasks, updates, or advice, they represent a transformative shift in work efficiency. This innovation highlights AI's evolving impact on work, providing dynamic and collaborative solutions that redefine traditional employment concepts.
AI in Medicine
Probably, one of the most nobel uses of AI is to be found in medicine. AI is making big strides in changing how we diagnose illnesses. In a UK study on breast cancer, using AI to read mammograms reduced the chances of getting things wrong by 5.7% for false positives and 9.4% for false negatives. In South Korea, another study found that AI was better at spotting breast cancer, especially in finding lumps (90% accuracy compared to 78% for radiologists) and catching early-stage cancer (91% accuracy compared to 74% for radiologists). Just think of how many lives could be saved if diagnoses are established in the early stages of disease development.
The Learnings
All of the examples above show how AI advancements can have ripple effects across various professions. But, what we need to understand is that AI won’t replace people doing their jobs - it will replace people who don’t use AI to do their jobs. AI is here to help us and it is up to us to take advantage of its power. The benefits are limitless.
As AI rapidly takes over various industries, the key factor shaping its impact is the data it relies on. What we feed into AI models determines their learning and, consequently, what unfolds in the real world. This crucial concept has the potential to make or break businesses and has prompted the establishment of new rules and regulations to steer AI's course.
The responsibility now rests on us—the custodians of this data-driven revolution. With the AI landscape evolving at a fast pace, there's a pressing question: should we hit the brakes and reassess amidst this speedy transformation?
As AI progresses, there's a shift from human-guided control to independent exploration. However, concerns arise, highlighted by an alarming incident where AI generated plans for 10,000 chemical weapons in just six hours. That is why the need for responsible AI development has been recognized, and global rules are now in place. In the U.S., President Biden's Executive Order sets new standards for AI safety, privacy, equity, and innovation. Similarly, the EU's AI Act categorizes systems by risk, prohibiting high-risk AI and requiring transparency for generative AI. It's a shared challenge to balance technological progress with ethics.
AI is exploding - no doubt about that. It is advancing incredibly, particularly with the powerful transformer architecture. Traditionally, machines learned from human-provided data, but now they're gaining the ability to learn on their own—a game-changing moment in the AI industry.
And, as we're on the brink of a revolution in the AI industry, businesses need to take charge and get ready for what's coming. Here are some suggestions:
1. Sense of Urgency: Businesses should recognize the need for swift AI adoption. Companies that embrace AI early on gain a competitive edge, becoming more agile and efficient. Boards of directors should develop strategic plans, empowering teams to explore AI applications for daily operations, driving innovation and process improvement.
2. Prioritize Ongoing Learning: Given the fast-paced nature of AI development, businesses should invest in continuous learning programs and AI awareness. This ensures the workforce stays adept at leveraging evolving AI technologies, adapting to new challenges, and maximizing the benefits of personalized AI solutions.
3. Invest in Responsible AI Practices: Businesses should approach AI development with a strong commitment to ethical practices. Everybody needs to understand the potential risks of unsupervised learning, learning from past instances like biased models (as seen in Amazon's hiring algorithm). We should implement safeguards to prevent unintended outcomes, ensuring fairness and inclusivity in AI applications.
In conclusion, the fusion of AI with the global industry showcases the tremendous impact technology can have on our world. Moving ahead, businesses need to step into the AI revolution with a mindful approach, being cautious, responsible, and dedicated to ethical practices. This way, we guarantee that AI continues to be a force for good, driving innovation and positive change while steering clear of potential risks.
The future is AI, and by journeying down this road alongside technology, we can craft a world where innovation, responsibility, and progress thrive together in unison.
Looking forward, I'm currently writing an article on specific applications of AI in e-commerce, so stay tuned for more insights directly from the core of the AI world.
* Originally posted on LinkedIn
Pixyle AI generates e-commerce product data that enables brands, retailers and marketplaces to deliver exceptional product discovery experiences.
