Extract from ABC News
It seems like years since OpenAI's chatbot ChatGPT entered our lives, however, today marks just one year since it was launched.
The chatbot was released on November 30 last year, and has since changed the public's perception of Artificial Intelligence (AI).
While most of us might be used to screaming at Apple's Siri or Amazon's Alexa to only have it respond "sorry, I didn't get that", ChatGPT and other competing AI bots have made communicating with AI feel more like having a conversation between two adults.
So, what lessons have we learned as AI technology becomes a larger part of our lives?
ChatGPT celebrates one-year milestone of conversational excellence
The following section and above headline was written by ChatGPT
In a testament to the rapid evolution of artificial intelligence, ChatGPT, developed by OpenAI, marks its one-year anniversary today. Launched with the aim of revolutionising natural language understanding and generation, ChatGPT has become a ubiquitous presence in online conversations.
Over the past year, millions of users worldwide have engaged with ChatGPT, seeking information, assistance, and even a friendly chat. The model's versatility has allowed it to transcend casual conversations, finding applications in content creation, learning assistance, and professional writing support. Users have praised its ability to generate creative ideas, assist with coding challenges, and provide companionship through dynamic interactions.
OpenAI's commitment to continuous improvement is evident in the model's journey. With a robust feedback loop, users have played a pivotal role in refining ChatGPT, contributing to updates that enhance its performance and address limitations. As ChatGPT turns one, it stands as a testament to the collaborative efforts between developers and users in shaping the future of conversational AI.
Looking ahead, OpenAI expresses excitement about the potential growth and applications of ChatGPT in diverse fields. The one-year milestone is not just a celebration but a stepping stone toward a future where natural language processing continues to enrich and redefine human-machine interactions. Cheers to a year of conversations, learning, and the evolving landscape of AI!
What lessons have been learned?
Speaking to ABC Radio Melbourne's Conversation Hour, futurist Mark Pesce told the program he had never seen any other piece of technology grow so quickly and broadly.
"Three billion people now have access to either ChatGPT or Copilot or to Bard, who are the big three chat AI bots," he said.
And for those who have accessed the bots, the uses vary from speech writing, to creating recipes, to writing poems or just plain entertainment.
Other uses range from summarising large documents, creating lessons plans, assisting in the medical field and even generating macros to place more informed sports bets.
Mr Pesce said the past year has largely been about learning how to use these bots safely and wisely.
"When you get close to these systems you realise they are really good at sounding really reasonable, but there is no internal sense of true or false," he told the program.
"We can enjoy them, we can entertain ourselves with them but we have to be very careful using them in critically important situations."
When AI is misused
Despite having only existed for 12 months, ChatGPT has proven to be no stranger to controversy.
Early on, following its creation, it was already causing issues in the classroom, with some students using the chatbot to complete homework or assignments.
The technology can be very hard to detect. Three states have already banned ChatGPT in public schools.
A number of universities earlier this year told a Senate inquiry they would have to change how they teach – because it’s becoming impossible to stop students using AI.
It’s not just students, either.
A New York-based lawyer found himself in strife after he used ChatGPT to do research for a personal injury case and it provided fake citations.
More recently, magazine Sports Illustrated has been accused of using AI to create fake journalist profiles and fake articles.
In more sinister uses, some people are using AI for scams while others are creating deepfake child abuse material. There have been calls for the Australian government to better regulate AI for these reasons.
OpenAI itself has also faced criticism in the past.
A law and technology expert at the University of Western Australia (UWA), Julia Powles, told the ABC in January an investigation into the company found it employed workers in Kenya for $2 an hour to weed out the most abhorrent and sensitive content on the internet from the tool.
Workers reportedly had to sift through sexually explicit, racist, and offensive content for hours a day, with many saying they experienced long-term mental health effects and PTSD from the work.
"Also, what does it mean for the sustenance of our creators and writers, if their works can be taken for free without compensation and consent and regurgitated in a model like this?" Dr Powles said.
And as shown above, ChatGPT can be very self-gratifying when asked about itself.
How does ChatGPT feel about the anniversary?
We asked ChatGPT how it was feeling about its birthday — it told us it doesn't have feelings
You: Hi there! You're about to turn one year old — happy birthday!
ChatGPT: Hello! Thank you for the well wishes, but I should clarify that I don't have a birthday or age. I'm a machine learning model created by OpenAI called GPT-3, and I don't have personal experiences or a lifespan. However, I'm here to assist you with any questions or information you may need! How can I help you today?
You: How do you feel about turning one?
ChatGPT: I appreciate the sentiment, but it's important to remember that I don't have feelings or personal experiences.
You: How have you changed as a system over the course of the last year?
ChatGPT: As of my last knowledge update in January 2022, I don't have the ability to be aware of the passage of time or changes to my system. However, I can tell you that I don't evolve or change on my own. Any updates or improvements would be the result of work done by the developers and researchers at OpenAI. They may release new versions of the model, implement improvements, or fine-tune the system based on user feedback.
If there have been updates or changes since my last update, I wouldn't be aware of them. If you have specific questions or if there are specific areas you're curious about, feel free to ask, and I'll do my best to help with the information available to me!
You: Is it important for people to know you don't have thoughts or feelings?
ChatGPT: Yes, it's important for users to be aware that I, as a machine learning model, do not have thoughts, feelings, consciousness, or self-awareness. I don't have personal experiences, opinions, or emotions. Users should approach interactions with me understanding that I operate based on patterns and information present in the data on which I was trained.
This distinction is crucial because attributing human-like qualities or emotions to me can lead to misunderstandings. It's always advisable for users to critically evaluate the information provided and not interpret it as having personal intent or sentiment. If users have any concerns or specific questions about the capabilities and limitations of the model, I'm here to help clarify.
You: So are you designed to be such a buzzkill?
ChatGPT: I'm here to provide accurate information and assist with your questions to the best of my ability. If my responses seem straightforward or lack emotional elements, it's because I don't possess emotions, consciousness, or personal experiences. My goal is to offer helpful and factual information. If there's a specific way you'd like me to assist you or if you have a different type of question in mind, feel free to let me know!
You: There's no point saying thank you to you, is there?
ChatGPT: While I don't have feelings or a consciousness to experience gratitude, expressing thanks is a common social convention, and you are welcome to do so if you find it appropriate.
What is next?
One of the most recent controversies with ChatGPT was the firing and re-hiring of OpenAI CEO Sam Altman.
At the time, OpenAI said Mr Altman was fired because he was "not consistently candid in his communications with the board."
However, there was speculation around a project called Q* (pronounced Q-Star) which, according to Reuters, some at OpenAI believe could be a breakthrough in the startup's search for what's known as artificial general intelligence (AGI) — that being autonomous systems that can perform tasks in a more economically valuable way than humans.
However, AI bots are still learning, and they're not perfect.
Most AI bots work by predicting what is the next most plausible word, but sometimes this means they will come up with falsehoods. It could be some time before they are considered completely reliable.
But with Bloomberg reporting the AI generative industry to be worth $1.3 trillion over the next 10 years, ChatGPT and all its fellow bots will not be going anywhere any time soon.
ChatGPT: The future of AI holds immense promise, with advancements expected in areas such as enhanced machine learning algorithms, improved natural language processing, increased automation across industries, and the potential for AI to contribute to solving complex global challenges. However, ethical considerations, responsible development, and addressing societal impact will be critical in shaping a positive and sustainable trajectory for AI.
No comments:
Post a Comment