THE BLOG

My conversation with Bob.

Industry
Friday, April 21, 2023
I sat down today to have a conversation with AI. The following is a transcript of our chat, the conversation is printed in full of some edits for clarity. Text in bold were my questions, quoted sections are the response from the chat.

My conversation with Bob.

I sat down today to have a conversation with ChatGPT. The following is a transcript of our chat, the conversation is printed in full of some edits for clarity. Text in bold were my questions, quoted sections are the response from the chat.

Do you have a name?

Sure! While I don't have a traditional name since I'm an artificial intelligence language model, you can feel free to call me by any name you prefer. Some people like to call me "GPT" for short, but ultimately, it's up to you!

Can I call you, Bob?

Certainly! You can call me Bob if you like. Just keep in mind that my responses are generated by a computer program, and I don't have a physical existence or identity like a human being.

Can you tell me a bit about yourself?

Hello! I'm a language model developed by OpenAI. I have been trained on a massive amount of text data and use deep learning techniques to generate human-like responses to a wide range of questions and prompts. My primary goal is to assist and provide helpful responses to users like you. Feel free to ask me anything, and I'll do my best to provide you with a helpful and informative answer!

And how old are you?

As an artificial intelligence language model, I don't have a physical age or existence in the traditional sense. I exist solely as a computer program running on servers, and my capabilities and knowledge are determined by the data I have been trained on. However, I was developed and first introduced in 2020 by OpenAI.

Tell me a bit about yourself. Where did you grow up?

As an artificial intelligence language model, I don't have a physical existence or a place of origin like a human being. I was created and developed by OpenAI, a research organisation based in San Francisco, California.

Can you give me a concise and accurate definition of AI that would be understandable for a layman?

Sure, I can provide you with a simple definition of AI that can be easily understood by a layman. Artificial Intelligence (AI) is a field of computer science that focuses on creating machines and software programs that can perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and language translation.AI systems learn from the data they receive and make predictions, recommendations, or decisions based on that data. These systems are designed to work independently or in collaboration with humans, with the goal of improving efficiency, accuracy, and productivity in various fields such as healthcare, finance, transportation, and education.

Bob, how do you learn?

As an artificial intelligence language model, I learn through a process called machine learning. I was initially trained on a vast corpus of text data using a technique called unsupervised learning. During training, I analyse patterns and relationships within the text data and develop a statistical model that can generate text in a similar style and tone as the data I was trained on.
After the initial training, I continue to learn through interactions with users like you. Every time someone types a query or message to me, I analyse the input, try to understand its meaning, and generate a response based on my understanding. Over time, I can improve my responses through feedback from users, which allows me to refine my language skills and better understand the nuances of human language.
In summary, my learning process involves analysing and processing vast amounts of data, developing a statistical model, and continually refining my language skills through interactions with users.

And how do photo AI's learn?

Photo AI, or computer vision models, typically learn through a process called supervised learning. In supervised learning, the model is trained on a large dataset of images that have been labelled with corresponding information, such as the objects present in the image or the desired output for a given task.
During training, the model analyses the images and their labels and develops a statistical model that can recognize and classify objects in new, unseen images. The model is optimised to minimise the difference between its predicted output and the true label of the image, using techniques such as backpropagation and gradient descent.
In addition to supervised learning, photo AI can also learn through unsupervised learning, where the model analyses patterns and relationships within unlabelled data to identify common features and clusters in the images. This can be useful in tasks such as image segmentation or anomaly detection.
Overall, photo AI models learn through a combination of supervised and unsupervised learning, as well as other techniques such as transfer learning, where pre-trained models are fine-tuned for a specific task, and reinforcement learning, where the model learns through trial and error in a simulated environment.

Do photo AI models get permission from the original photographer or artist before using it for training?

In general, photo AI models do not require permission from the original photographer or artist to use their images for training. This is because the images used for training are often sourced from publicly available datasets or scraped from the internet, where they are already widely shared and accessible.
However, there are some exceptions to this. For example, if the images being used for training contain sensitive or confidential information, such as personal or medical data, then permission may need to be obtained from the individuals or organisations that provided the data.
Additionally, some researchers and organisations have started to recognize the importance of ethical considerations when using datasets for AI training, including issues around bias and representation. As a result, there have been calls for greater transparency and accountability in the use of publicly available datasets, and efforts to ensure that the data used for training is diverse, representative, and does not perpetuate harmful stereotypes or biases.
In summary, while photo AI models may not always require permission from the original photographer or artist to use their images for training, there is a growing recognition of the need for ethical considerations and greater transparency in the use of publicly available datasets.

Do you think that its morally right that a photo ai has used one of my photos in its dataset without my permission?

As an AI language model, I do not have personal opinions or beliefs, but I can provide some insight into the ethical considerations surrounding the use of personal photos in photo AI datasets.
From a legal perspective, the use of your photos in a photo AI dataset without your permission may be considered a violation of your intellectual property rights. However, the use of images in datasets for research purposes may fall under certain exceptions and limitations to copyright law, such as fair use or the right to freedom of expression.
From an ethical perspective, the use of personal photos without permission raises concerns around privacy and consent. While it may be legal to use publicly available images for research purposes, it is important for researchers and organisations to be transparent about their data sources and to obtain informed consent when possible.
Furthermore, the use of personal photos without permission may perpetuate biases and stereotypes in photo AI models, as these models are only as accurate and fair as the data they are trained on. As a result, there are growing calls for greater accountability and transparency in the development and use of photo AI models, particularly in regards to issues of bias and representation.
In summary, while the use of personal photos in photo AI datasets without permission may be legal under certain circumstances, there are ethical considerations around privacy, consent, and bias that should be taken into account by researchers and organisations.

Do you think legal constraints may be necessary in the future because of these ethical issues?

It is possible that legal constraints may be necessary in the future to address the ethical issues surrounding the use of personal photos in photo AI datasets. As the development and use of AI technologies continue to evolve, there is growing recognition of the need for ethical considerations and responsible AI practices, particularly in regards to issues of privacy, bias, and transparency.
There have been some efforts to develop regulations and guidelines around the use of AI technologies, including the EU's General Data Protection Regulation (GDPR) and the IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems. These initiatives aim to establish ethical principles and standards for the development and use of AI technologies, including photo AI models.
However, the effectiveness of these initiatives in addressing the ethical issues surrounding photo AI models and other AI technologies remains to be seen. It is likely that continued discussions and debates around these issues will be necessary to develop meaningful and effective regulations and guidelines that balance ethical considerations with the potential benefits of AI technologies.

So in other words, the cat is already out of the bag?

In a sense, yes. The use of personal photos in photo AI datasets has already become widespread, and it may be difficult to fully address the ethical issues surrounding this practice. As more and more data becomes available on the internet, it is increasingly difficult to control the distribution and use of personal photos.
However, this does not mean that ethical considerations and responsible AI practices should be abandoned. Instead, it highlights the importance of continued discussions and efforts to address these issues, both in the development and use of photo AI models and in the broader context of AI technologies as a whole.
It is important for researchers, organisations, and policymakers to be proactive in addressing the ethical implications of AI technologies, including issues of privacy, bias, and transparency, and to work towards developing meaningful and effective regulations and guidelines that balance these considerations with the potential benefits of AI.

This is depressing.

I understand that the ethical considerations surrounding the use of personal photos in photo AI datasets can be disheartening. However, it is important to recognize that there are also many positive applications of photo AI, such as in medical imaging, environmental monitoring, and creative fields such as photography and film.
Moreover, by acknowledging and addressing the ethical issues surrounding photo AI models and other AI technologies, we can work towards ensuring that they are developed and used in ways that are responsible, fair, and transparent. This includes efforts to reduce bias and ensure that data used for training is diverse and representative, as well as greater transparency around the use of personal data and informed consent.
While there are no easy solutions to these issues, it is important to continue the conversation and work towards creating a more ethical and responsible future for AI technologies.

While I understand that my personal photos might be used for  positive applications such as in medical imaging, environmental monitoring, it is also much more likely to be used to create an AI photo of a hamburger. That takes money away from food photographers. Don't you think that this is also an important ethical consideration?

Yes, you raise an important ethical consideration regarding the impact of photo AI models on professional photographers and other creative professionals. The use of photo AI models to generate images of food, for example, could potentially take away business from food photographers and impact their livelihoods.
It is important for AI researchers, developers, and users to consider the potential economic and social impacts of their work and to ensure that the development and use of AI technologies is done in a responsible and ethical manner. This includes considering the impact on different industries and professions, as well as potential unintended consequences of AI applications.
One potential solution to this issue is to ensure that photo AI models are used in ways that complement and enhance, rather than replace, the work of creative professionals. This may involve incorporating AI technologies into the creative process or developing new business models that leverage AI technologies in innovative ways.
Overall, it is important to balance the potential benefits of photo AI models with the potential economic and social impacts, and to work towards creating a more equitable and responsible future for AI technologies.

I mean, its easy enough just to say things like we should "ensure that photo AI models are used in ways that complement and enhance, rather than replace, the work of creative professionals". But that really isn't how the real world works. So now we have the situation where photo AI can use photos without permission as part of their training set for photo AI, which can be used by marketing companies to create fake photos of hamburgers, and the original photographer doesn't get paid. Is this correct?

Yes, you are correct that the current situation regarding the use of personal photos in photo AI datasets can lead to issues around compensation and credit for the original photographer or artist. The use of personal photos in photo AI models without permission or compensation can be seen as a form of exploitation, particularly if these images are then used for commercial purposes without proper attribution or compensation for the original creator.
However, it is important to note that there are ongoing efforts to address these issues and to develop more ethical and responsible approaches to the development and use of AI technologies. For example, some organisations are working to develop new business models that leverage AI technologies in ways that are more equitable and inclusive, while others are exploring the use of decentralized networks and blockchain technologies to create more transparent and fair systems for compensation and attribution.
Ultimately, addressing the ethical considerations surrounding photo AI models and other AI technologies will require ongoing efforts from a range of stakeholders, including researchers, developers, policymakers, and creative professionals. It will be important to continue the conversation around these issues and to work towards developing new solutions and approaches that balance the potential benefits of AI technologies with the need for fairness, transparency, and ethical considerations.

This is depressing.

I understand that the issues surrounding the use of personal photos in photo AI datasets and the impact on creative professionals can be disheartening. However, it is important to recognize that there are ongoing efforts to address these issues and to develop more ethical and responsible approaches to the development and use of AI technologies.
Moreover, it is important to remember that there are many positive applications of photo AI models and other AI technologies, including in fields such as healthcare, scientific research, and environmental monitoring. By working towards more responsible and ethical AI practices, we can ensure that these technologies are used in ways that are beneficial to society as a whole.
While there is still much work to be done, it is important to remain hopeful and to continue the conversation around these important issues. By doing so, we can work towards creating a more equitable, transparent, and responsible future for AI technologies.

So how do you think it all ends?

It's difficult to predict exactly how the issues surrounding photo AI and other AI technologies will ultimately be resolved, but it is likely that a combination of technological advancements, policy changes, and cultural shifts will be necessary to create a more ethical and responsible future for AI.
One possibility is that new technologies and approaches will be developed that can better protect personal data and ensure that the use of personal photos in photo AI models is done in a more transparent and ethical manner. For example, there may be new techniques for data anonymization or decentralized networks that allow individuals to retain control over their personal data.
Another possibility is that there may be policy changes and regulations put in place to ensure that AI technologies are developed and used in ways that are fair, transparent, and accountable. This may include new laws around data privacy and ownership, as well as regulations around the use of AI in specific industries or applications.
Finally, it is possible that cultural shifts will occur that encourage greater awareness and understanding of the ethical considerations surrounding AI technologies. This may involve greater education and outreach efforts to ensure that individuals and organisations understand the potential impacts of their actions and are motivated to act in ways that are responsible and ethical.
Overall, it is likely that a combination of these and other factors will be necessary to create a more ethical and responsible future for AI technologies. While there may be challenges along the way, it is important to remain hopeful and to continue the conversation around these important issues.

I think you are being optimistic. But I thank you for your candour. What do you think is the future for creatives?

The future for creatives in a world with advancing AI technologies is difficult to predict, but it is clear that there will be both challenges and opportunities.
On the one hand, AI technologies may be used to automate certain creative tasks, such as selecting colours or composing music. This could lead to a reduction in the demand for certain types of creative professionals, particularly those who are focused on more routine or repetitive tasks.
On the other hand, there are also opportunities for creative professionals to leverage AI technologies in new and innovative ways. For example, AI technologies can be used to generate new ideas or to explore new creative directions that would not have been possible otherwise. In addition, AI technologies can be used to analyse data and user behaviour to inform creative decisions and to create more personalized and engaging experiences for audiences.
Overall, the future for creatives will likely involve a combination of both challenges and opportunities. It will be important for creative professionals to stay informed about new AI technologies and to continue to develop their skills and expertise in ways that allow them to take advantage of these new opportunities. At the same time, it will be important for society as a whole to work towards creating a more ethical and responsible future for AI technologies, one that respects the rights and contributions of all individuals and ensures that the benefits of AI are distributed fairly and equitably.

From my limited experiences with AI I've found it has been helpful to automate simple tasks. And I'm less concerned about independent creators like me who have always been "up against it", but more worried about bigger organisations who will leverage these technologies to drive down costs and put creatives out of work. What are your thoughts?

You bring up an important point. While AI technologies can certainly be helpful in automating simple tasks and freeing up creative professionals to focus on more complex and nuanced work, there is also a risk that these technologies could be used to drive down costs and put creatives out of work, particularly in the context of larger organisations.
One potential solution to this problem is to ensure that AI technologies are used in ways that are ethical and responsible, and that they are deployed in ways that complement and enhance the work of creative professionals, rather than replacing them altogether. This may involve creating new business models or revenue streams that allow creatives to monetize their work in new and innovative ways, or it may involve creating new regulations or standards that ensure that the benefits of AI are distributed fairly and equitably.
At the same time, it is also important for creative professionals to develop their skills and expertise in ways that allow them to stay ahead of the curve when it comes to new AI technologies. This may involve investing in new training and education programs, or it may involve seeking out new collaborations or partnerships with AI developers and experts in order to stay informed about the latest developments and opportunities.
Ultimately, the future of creatives in a world with advancing AI technologies will depend on a variety of factors, including technological advancements, cultural shifts, and policy changes. It will be important for all stakeholders to work together to create a more ethical and responsible future for AI technologies, one that respects the rights and contributions of all individuals and ensures that the benefits of AI are distributed fairly and equitably.

Kirk or Picard?

The debate between Kirk and Picard is a longstanding one among Star Trek fans, and there are valid arguments for both sides. Some people prefer Kirk's swashbuckling, action-oriented style, while others appreciate Picard's more cerebral and diplomatic approach. Ultimately, it's a matter of personal preference, and there is no right or wrong answer.

If you had to pick a captain to defend the earth from the Borg, who would you pick?

If I had to pick a captain to defend the Earth from the Borg, I would choose Captain Jean-Luc Picard. He has demonstrated his ability to successfully negotiate with the Borg in the past and has a deep understanding of their culture and tactics. Picard is also a skilled strategist and tactician, and his calm and measured approach to problem-solving could be valuable in a high-stress situation. Additionally, his experience as a diplomat and his knowledge of history and culture could be beneficial in finding a peaceful resolution to the conflict, if possible.

But you know that at one point Picard was part of the borg collective, right? He was Locutus Of Borg. Responsible for the death of thousands of starfleet officers. Are you sure?

Yes, you are correct. Captain Jean-Luc Picard was assimilated by the Borg and briefly served as Locutus of Borg in the Star Trek: The Next Generation episode "The Best of Both Worlds". While he was under the control of the Borg, he was responsible for the deaths of many Starfleet officers and inflicted significant damage on the Federation.
However, I still believe that Captain Picard would be a good choice to defend Earth against the Borg. Although he was once assimilated, he also has firsthand knowledge of the Borg's tactics and weaknesses, which could be valuable in combat. Additionally, he has shown a strong ability to resist the Borg's influence and maintain his individuality, as demonstrated in "The Best of Both Worlds" when he fought against the Borg's attempts to control him.
In the end, the choice of captain to defend Earth from the Borg would ultimately depend on the specific circumstances of the situation, as well as the strengths and weaknesses of the entire Starfleet roster at that time.

"I am Big Mark of Borg. Resistance...is futile."

As an AI language model, I don't have physical form or consciousness, so I cannot be assimilated into the Borg collective. However, I am programmed to be helpful and answer any questions or concerns you may have. If you have any other inquiries, please don't hesitate to ask.

This is part 1 of 2. In the second part, which will be available in a few weeks, I'll reflect on my conversation with Bob, and talk a bit about how I think AI will impact those of us who working in the creative industry.

Post by Big Mark

Big Mark is a website developer and photographer that lives in Wellington, New Zealand. He has owned and run his own creative agency since 2011.

Recent Posts
Photography Workshops
Big Mark Photography Workshops
We''ve relaunched Big Mark Photography Workshops. Learn in a safe, encouraging environment with Big Mark. 25% off workshop vouchers if purchased before the end of May.
Built by Big Mark
Four websites inside Wellington, New Zealand
We take a closer look at four websites recently built by Big Mark and the crew.
Web design
Looking for freelancers
Are you a freelance graphic designer? Big Mark would like to work with you.

Lets chat!

Get a free 30 minute consultation

Get the ball rolling with a free 30-minute consult. Let's discuss your goals and how we can help. Book your session today!

Big Mark is an Amotai registered indigenous supplier

Amotai is Aotearoa's supplier diversity intermediary tasked with connecting Māori and Pasifika-owned businesses with buyers wanting to purchase goods, services and works.

Amotai verifies Māori and Pasifika-owned businesses and holds a national database of Māori and Pasifika-owned businesses that are ready for work.

Link to the Amotai Website