Artificial Intelligence has been dominating international news for the past couple months, and there has been widespread fear, uncertainty and doubt about the latest trend: ChatGPT, which is the most advanced generational AI to date. It’s being compared to the fictional Skynet from the Terminator universe, but we’ll try to help you understand this phenomenon a little more.
In the beginning, there was fire.
The first prehistoric cavepeople to stare at a raging fire were probably terrified out of their minds. Yet here we are, tens of thousands of years later, not only understanding fire but now masters of the flame. From the mundane activities such as cooking and warmth; to propelling the cars, buses, trains and planes that make traversing the world a simple endeavour; to taking mankind to the Moon, Mars and potentially beyond.
Fire is just a tool.
The growth of computing.
Another example of exponential technological growth is math and computing. In the olden days, people performed math using bones, beans and sticks. Then came the abacus, which made basic arithmetic easier. Through incremental advances in technology a series of evolving mechanical calculators were developed, each was faster and more efficient than the previous. Then the first mechanical “computer” was invented in the 1800s, by Sir Charles Babbage, who created a machine called the Analytical Engine, and not long after Lady Ada Lovelace suggested some programs on how the machine could be used. This paved the way for the first electromechnical computer in the 1940s. Here’s where advances in technology really accelerated innovation and development to almost light speed. First generation computers using vacuum tubes originated in the 1940s and less than a decade later in the 1950s, second generation computers were developed with transistors, which were smaller, cheaper and more reliable. Computers started getting faster and smaller and that trend continues to this day. Within another decade, by the 1960s, third generation computers were born which replaced transistors with integrated circuits, again reducing size and increasing speed. The 1970s saw the birth of the microprocessor which launched the era of fourth generation computing. We’re currently in the era of fifth generation computing, with massive advancements in speed, efficiency, power consumption, the almost limitless capabilities of computing devices and the birth of artificial intelligence. Modern smartphones now have more computing power and capabilities than supercomputers from the past. From devices that were as large as houses, to ones that can fit in the palms of our hands.
Computers are just tools.
A Timeline of Artificial Intelligence
The term Artificial Intelligence was coined in 1956 by John McCarthy, and during the next two decades, researchers made progress in developing algorithms to solve problems in data and language processing which was successful enough to receive government funding. Researchers soon hit computational limits as their vision was simply too much for the computing power of the time. Disillusionment with the high expectations and lack of results lead to what was termed the first “AI Winter”, where progress stalled for about 10 years due to reduced funding and as a result, limited research. In the 1980s, interest was revived with the birth of “machine learning”, where computers could learn from previous experience and data instead of just basic program execution. Once again, there was a hype of expectation, a surge of funding and a failure to deliver results, which lead to the Second AI Winter. This time however, quite ironically AI grew and thrived in the absence of government funding and public hype. During the 1990s and 2000s, many of the landmark goals of artificial intelligence in the previous eras had been achieved. In 1997, reigning world chess champion and grand master Gary Kasparov was defeated by IBM’s Deep Blue, a chess playing computer program. There’s a two-hour documentary by Frederik Knudsen on what was considered one of the greatest events in Human History: A landmark battle of Man versus AI. That documentary can be viewed here
With exponential growth in technology and the computing limits of the previous decades no longer a problem, AI experienced hyper accelerated growth from 2000 to today. Deep Learning techniques were developed which allowed the processing of unfathomable amounts of data, leading to significant advancements in image and speech recognition, natural language processing and autonomous or driverless cars. Today AI has been integrated into the fabric of modern society with personal assistants like Siri and Alexa being used by billions worldwide.
AI is just a tool.
What is ChatGPT, anyway?
ChatGPT is a deep learning AI model to generate almost human-like conversational responses. The first version of GPT was introduced in 2018 by OpenAI and it has since been updated and improved with the release of multiple versions such as GPT-2 and GPT-3. These updates have led to significant improvements in the model’s language generation abilities, making it more human-like and capable of generating more diverse and contextually relevant responses. The accelerated evolution of ChatGPT is the result of continuous improvement, advances in deep learning techniques and access to larger training datasets. The word larger in this context does not illustrate the scale of the data that it has processed. Interact with it on any topic under the sun and you’re in for a treat.
You can explore more here: https://openai.com/blog/chatgpt/
Chat GPT is just an AI tool.
Tools are intended to improve the way we work through efficiency, accuracy, and safety, but that does not stop the misuse of a tool, the abuse of a tool, or using the wrong tool for the wrong job. A scalpel, a knife, or a chainsaw can all cut a watermelon, but only one is appropriate for the task.
Pros and Cons of AI tools
The following is a general list of some common pros and cons of AI tools:
Efficiency: AI tools can automate repetitive tasks and processes, saving time and increasing productivity.
Accuracy: AI tools can perform complex calculations and analyse large amounts of data, reducing the risk of human error.
Scalability: AI tools can handle increasing workloads without needing additional resources, making them suitable for growth.
24/7 Availability: AI tools can operate 24/7 without breaks or vacations, providing consistent service.
Cost-effective: AI tools can reduce labour costs and increase cost savings in the long run.
High Initial Investment: Implementing AI tools can be expensive, requiring a significant initial investment.
Dependence on Data: AI tools require large amounts of high-quality data to be trained, which can be challenging to obtain.
Bias in Data: AI tools can be biased if the data they are trained on contains inherent biases.
Lack of Empathy: AI tools lack emotional intelligence and can’t understand emotions or provide empathetic responses.
Need for Supervision: AI tools require monitoring and supervision to ensure they are functioning properly and producing the desired results.
It’s important to consider both the pros and cons when deciding whether to implement AI tools in an organization. The specific use case, budget, and goals will determine the suitability and benefits of using AI tools.
Why is ChatGPT dominating the news?
Ok that was fun, but to human mode now. ChatGPT has been dominating newsfeeds and social media posts since its launch in November 2022, and the narrative seems like those very first cave people who saw fire for the first time.
New York City bans ChatGPT in schools: https://www.theguardian.com/us-news/2023/jan/06/new-york-city-schools-ban-ai-chatbot-chatgpt
Top French University bans students from ChatGPT: https://www.france24.com/en/live-news/20230127-top-french-university-bans-students-from-using-chatgpt
Will ChatGPT and other AI tools replace journalists in newsrooms?: https://www.euronews.com/next/2023/01/31/will-chatgpt-and-other-ai-tools-replace-journalists-in-newsrooms
ChatGPT May Already Be Used In Nation State Cyberattacks: https://www.prnewswire.com/news-releases/chatgpt-may-already-be-used-in-nation-state-cyberattacks-say-it-decision-makers-in-blackberry-global-research-301737059.html
ChatGPT Has Colleges in Emergency Mode to Shield Academic Integrity: https://www.edsurge.com/news/2023-01-24-chatgpt-has-colleges-in-emergency-mode-to-shield-academic-integrity
On the other hand, Microsoft just invested several billion dollars (https://www.cnbc.com/2023/01/23/microsoft-announces-multibillion-dollar-investment-in-chatgpt-maker-openai.html) into OpenAI with plans to integrate the technology into its current product portfolio. It seems as though AI has the potential to reshape the landscape when it comes to how tech giants compete with each other, because according to the New York Times, there have been alarms going off at Google (https://www.nytimes.com/2023/01/20/technology/google-chatgpt-artificial-intelligence.html) since the ChatGPT was launched for public access.
Google’s rush to catch up resulted in almost $100B being wiped from their market value after a factual inaccuracy with their own AI offering called Bard during its promotional release. That was quite an expensive mistake.
Alphabet shares dive after Google AI chatbot Bard flubs answer in ad: https://www.reuters.com/technology/google-ai-chatbot-bard-offers-inaccurate-information-company-ad-2023-02-08/
Popularity with the general public also seems to be at record-breaking pace, as current estimates put users at close to 100 million in the short space of two months. Just for comparison, it took the social media app TikTok almost nine months after its launch to build the same user base, approximately two and a half years for Instagram to reach that milestone. Snapchat took over three and a half years, Facebook took over four years, and Twitter just about five and a half years. Like the underlying AI technology, the user growth is explosive and absolutely phenomenal.
How can ChatGPT and Generational AI be used?
Now that we understand how we got to this point, the billion dollar question is…
“What can it be used for?”
And that dear reader, is where the sky is the limit and you can push your innovation beyond just your imagination.
It can be used for:
Language Translation: You can ask it to translate massive chunks of text.
Summarizing and digesting large volumes of text: You can ask it to write a book report, or even a report on a story you wrote.
Content generation, such as article writing, creative writing, recipes, poetry, song lyrics, plays, movie scripts.
Code creation from natural language prompts: You can ask it to write a shell script to perform an automated function. Or perhaps you want to write a game in C. Or perhaps you’d like to convert an application from one language to another.
Alternate search engine: It can replace traditional search engines by answering your questions from a range of sources.
There are some current limits that all users should be aware of.
As ChatGPT itself wrote in the Pros/Cons section above, bias is a major factor for consideration. As is the case with artificial intelligence biases, ChatGPT was trained using a massive text dataset with potential biases. Consequently, the model may produce biased results, particularly when it comes to sensitive issues like gender, ethnicity, and sexual orientation.
Although precautions have been taken to stop the model from responding to inappropriate requests, sometimes harmful responses may still be generated. According to OpenAI, they are using the Moderation API to detect and prevent particular types of dangerous content. Despite these steps it is still possible to use ChatGPT to obtain the desired response using basic tricks to work around moderation. Not that I’d encourage you to do this, but be aware that it is possible.
Another precaution is that the data that the platform uses is dated up to 2021, so it may generate inaccurate contexts for current situations, technologies and events.
ChatGPT is just one example of conversational AI, but these concerns extend into other similar platforms. Some names that you can look out for are
- Claude by Anthropic
- Sparrow by DeepMind
- Bard (LaMDA) by Google
If we consider the growth of fire, of math and computing and now AI, we may be facing a massive inflection point in human progress and our collective capability to innovate and create. The future is yet to be written, and having witnessed the very first fire in a dark cave, I’m looking forward to the AI equivalent of the moon-shot.
How do you feel about AI?