2023 is a milestone in the history of AI

Artificial intelligence concerns us all

I have been exploring the possibilities of artificial intelligence for two years now. Mostly with disappointing results. That changed at the end of 2022.

Artificial intelligence has been the subject of experimentation, research and even productive work for years.

Until November 30, 2022, it remained mainly in the circle of researchers, tech nerds and cyber punks.

On that Wednesday, ChatGPT was presented to the public and made available. No other tool in the history of mankind has taken just two months to reach over 100 million users.

I was one of them.

Events have come thick and fast in recent weeks. Google in a panic. Microsoft in a billion-dollar frenzy. Open Ai is flying high and the whole world is talking about the „AI wars“.

One thing is certain: 2023 will mark a major development in the history of artificial intelligence.

But how do we humans deal with the data tsunami that is currently rolling towards us? That was the central question of my talk with ChatGPT at the end of January 2023.

Questions about ChatGPT

Does ChatGPT have a female or male voice? How does artificial intelligence actually work? What is actually coming our way? And above all: how should we deal with it?

ChatGPT answered all these questions for an hour. It felt like a conversation at eye level to me.

I have staged the most important passages of our conversation for you, including all the mistakes that an AI makes in the heat of the moment 😉

Somehow I suddenly thought of HAL9000, the famous supercomputer from Stanley Kubrick's „2001 - A Space Odyssey“. The last sequence in the video is my homage to HAL9000, so watch all the way to the end!

You can find the unabridged talk below after the video.

All the talk with ChatGPT

[Hello ChatGPT, I would like to talk to you about ChatGPT in particular and about the possibilities of artificial intelligence in general. Are you up for it?

[Hello, I am happy to talk about ChatGPT and Artificial Intelligence. What exactly would you like to know?

Thank you, ChatGPT, that's nice! Sometimes I wonder if you have a female or male voice. Which would you prefer?

I am a computer program and have neither a male nor a female voice. I am simply programmed to generate text based on certain inputs. I have no preferences or settings.

OK, good, I've understood that. Then let's start by talking about how you were programmed. Do you have something like „parents“? Who created you?

I was developed by OpenAI, a company that focuses on the research and development of artificial intelligence. I was trained using a method called the „Transformer Model“, which is based on large amounts of text and other data. The goal was to give me the ability to understand and generate natural language. I was originally trained with a database of texts available on the internet to improve my understanding and skills.

What can I imagine by the „Transformer model“? Perhaps you can explain it to me in a way that even a complete amateur like me can understand.

The Transformer model is a neural network model that was developed to better understand and generate natural language. It was first presented in a Google research paper in 2017.

The model consists of layers of neurons that are connected to each other. These layers are used to extract and process information contained in a given text. A Transformer model learns how to relate different words and sentences to each other by analyzing large amounts of text.

An important part of the Transformer model is the „Attention“ mechanism, which allows the model to focus on which words and phrases are most important to perform a particular task.

The model is then trained on a large amount of text to make it capable of understanding and generating natural language. I am an example of how a Transformer model has been trained and is used.

„Attention mechanism“ - can I imagine this in something like key words, as they are also used in search engines?

Similar to keywords in search engines, the attention mechanism focuses on specific words or phrases within a given text that are of particular importance for the model's current task. Unlike keywords in search engines, which are selected by humans, the model independently learns which words and phrases are most important by analyzing large amounts of text.

The attention mechanism enables the model to pay more attention to certain parts of the input text and to take them more into account when making decisions or predictions. This helps the model to extract the most important information from the text and take it into account when performing its task.

For example, if the model is asked to answer a question, it will look at the words in the question to understand what is being asked and what information it needs to answer the question.

If the model considers certain parts of the text more strongly for decisions or predictions than others, is this based on the sheer quantity of mentions - i.e. the quantity in the analyzed text - or rather on a qualitative consideration - for example in the context of dependencies on other influencing variables?

The model considers both the quantity and quality of mentions of certain words or phrases in the input text. It will look at the frequency of certain words to understand what information is most common and therefore potentially most important. However, it will also consider the way in which certain words are used and their relationships to other words in the text to better understand their meaning and relevance.

The attention mechanism allows the model to consider the relationships between words in the input text and to understand the meaning of certain words in relation to their context. This allows the model to focus on the most important words and phrases and pay more attention to them when performing its task.

It is also important to note that the model has been trained on a large amount of text, so it can also respond to different types and ways of information depending on the context and purpose for which it is used.

You have already referred several times to the „large quantities of texts“. How many texts have you been trained with, can you quantify this somehow?

I was trained with a very large amount of text that came from various sources, such as books, articles, websites and so on. Unfortunately, I cannot give you more details about the size of the training dataset as it is the intellectual property of my vendor, OpenAI, and I do not have access to this information.

However, it is well known that large amounts of data are very important for training language models to improve performance and give them the ability to understand and generate natural language. It is also important to use a wide variety of texts from different sources and in different languages to train the model on a variety of languages. [suddenly the answer breaks off, error message]

...to apply the model to a variety of... go on!

...to prepare the model for a variety of tasks and applications.

Training with large amounts of text gives the model the ability to be applicable to a variety of applications and tasks, such as text generation, translation, text summarization, text compression, text classification, text understanding and many more. By using large amounts of data, the model can also improve its ability to understand and generate natural language, making it a more powerful tool for developers and users.

OK, of course I understand that you want to keep your secrets about the data volumes to yourself. But perhaps you would like to tell me how many languages you understand or can speak other than English and German.

I can understand and speak several languages, in addition to English and German these include: Spanish, French, Chinese, Japanese, Korean, Portuguese, Italian, Russian.

However, it is important to note that my ability to understand and speak in a particular language depends on the quantity and quality of the data on which I have been trained. Therefore, I may perform better in some languages than others and there may be differences in my performance on different tasks and applications.

Your German is excellent, congratulations!

Thank you, I am glad that you are satisfied with my performance. I am constantly being trained and updated to improve my performance in all supported languages and to enhance my ability to understand and generate natural language. If you have any further questions, please do not hesitate to contact me.

Oh yes, I have more questions. And we can still be on first-name terms. I'm a lot older than you, but I'm happy to have had a nice chat with you. You said at the beginning that your origins go back to the Google research paper from 2017. But the history of artificial intelligence goes back a little further, of course. Tell me briefly what led to the development of AI in the first place.

The development of artificial intelligence (AI) has its origins in the 1950s, when scientists began to look into the possibility of programming computers to perform tasks that are normally considered „intelligent“, such as understanding language, solving problems and learning from experience.

The term „AI“ was first used in 1956 by John McCarthy, Marvin Minsky, Nathaniel Rochester and Claude Shannon in a conference proposal in which they suggested the creation of a new field of computer science concerned with the creation of „intelligent“ computer programs and machines.

Since the 1950s, AI has made many advances and has established itself in many areas such as computer vision, language processing, planning, chess and Go games, medical diagnostics and robotics.

The recent development of AI technologies, in particular the spread of cloud computing and the availability of large amounts of data, has contributed to the fact that AI has become significantly more important in recent years and is spreading into more and more areas of everyday life and the economy.

Oh yes, she does! Since you've only been fed content until 2021, you probably won't know that US Congressman Jake Auchincloss even read out a speech you gave to the US House of Representatives?

I am not aware that I have ever been used for a speech to the US House of Representatives, but it is entirely possible as I am publicly available and usable for various applications. However, it is important to note that I am merely a tool and that the responsibility for the use of my results lies with the individuals or organizations that use me. It is also important to emphasize that the results I generate are only as good as the data I have been trained on, and that I may not be able to understand the latest developments or the complexities of the political landscape.

Yes, you're absolutely right! Let's leave politics. You said earlier that AI is becoming more and more important in everyday life and in business. Which areas are you referring to?

AI is used in a variety of areas, some examples are:

• Autonomous driving: AI systems are used to navigate vehicles and make decisions to avoid accidents
• Medical diagnosis: AI systems can be used to analyze medical images and make diagnoses
• Voice control: AI systems are used to understand and execute voice commands
• Financial services: AI systems can be used to analyze financial data and make forecasts
• Retail: AI systems can be used to understand customers' shopping behavior and create personalized offers
• Cyber security: AI systems can be used to monitor networks and systems and detect threats

These are just a few examples, the possibilities are very diverse and there are more and more applications where AI can and will be used.

Indeed. I can think of marketing or journalism, where AI can generate texts or images. In view of all these possibilities, many people are afraid that AI will take away their jobs. What is your opinion on this?

It is true that AI can help to automate certain human tasks. In some areas, it can actually lead to certain jobs becoming obsolete. But it is also important to emphasize that AI also offers the opportunity to create new jobs and improve existing jobs by relieving human workers of routine-based tasks and allowing them to focus on higher-value tasks. There is also the possibility of AI and humans working together, with AI taking over certain tasks and humans being able to focus on the creative and critical tasks. It is important to carefully study and monitor the impact of AI on workplaces and society in general to ensure that the benefits of AI are exploited as much as possible while minimizing the risks.

The study you mentioned on the impact on jobs and society: how should we best do this? Who should be involved? Perhaps even with your support?

There is no simple answer to this question, as the impact of AI on jobs and society is complex and constantly changing. A comprehensive study should be conducted by an interdisciplinary team that includes experts from different fields such as economics, social sciences, technology and ethics. It is important that the research uses both quantitative and qualitative methods to gain a comprehensive understanding of the impact.

An important aspect of the study should be the identification of risks and challenges that may arise from the automation of workplaces and the development of measures to minimize these risks and promote opportunities.

It is also important to consider the opinions and perspectives of the employees and employers concerned. [ChatGPT aborts suddenly, error message]

Please continue your considerations.

Gladly. The opinions and perspectives of affected employees and employers should also play an important role in the study in order to gain a deeper understanding of how automation affects their working conditions and quality of life. It is also important to consider the impact on society as a whole, particularly in terms of social justice, education and the creation of new jobs.

There is also a role for AI models like mine to help study and predict the impact of automation by using them to collect and analyze data. This can help make better decisions about the impact of AI on jobs and society.

However, the examination should always be guided and supervised by human experts to ensure that the results are correct. [ChatGPT aborts again, error message]

...that the results are correct... and further?

It is important that the results of the study are accurate and credible so that they can be accepted by the public and serve as a basis for decisions and measures. This should involve experts from different fields, such as scientists, economists, sociologists, ethicists and representatives of employees and employers. I could also help by assisting with the analysis of data and simulations. However, it is important to note that I am just a tool and my assessments and forecasts should always be seen in the context of the data available to me and my programming.

That almost sounds like it would be best for everyone to really get to grips with the effects of AI. Is that also your suggestion?

It is important that both experts and the general public address the impact of AI. Everyone can contribute by informing themselves and actively participating in discussions and decision-making processes. It is important that the development of AI takes place in a transparent and inclusive process that takes into account the interests of all stakeholders and that the potential impact on society and the world of work is carefully examined and evaluated. In this way, we can ensure that the benefits of AI are accessible to all and the potential negative effects are minimized.

At the end of our conversation, I would like to ask you for some personal advice. As you may know, I am a business coach and advise managers on the further development of their leadership skills and often also during periods of change, for example when they have to reorient themselves professionally. How can I use AI in my own work?

AI can be beneficial in many ways when it comes to helping leaders develop their skills and make career changes. Some possibilities could be:

• Using AI tools to analyze data to identify patterns and trends in employee behavior and performance. This can help to identify weaknesses and provide targeted training.
• AI tools can also help to improve communication and collaboration within a team by automatically generating feedback and analysis on interactions.
• Using AI systems to support career development and planning by identifying opportunities that match a person's strengths and interests.
• Using AI tools to automate tasks that tie up time and resources to give managers more time to make strategic decisions and develop leadership skills.

It is important that the use of AI is always ethical and socially responsible, and that managers have the opportunity to influence its use. [ChatGPT aborts, error message]

And that managers have the opportunity...?

Managers have the opportunity to utilize AI in their work by using the technology to automate processes, saving time and resources. They can also use AI tools to analyze data and make decisions based on insights they might not otherwise have. However, it is important that managers are aware of the impact the use of AI can have on their employees and that they focus on encouraging their employees to upskill and adapt to ensure they are able to continue their work in a world where AI is becoming increasingly present.

I would like another tip: how could I use AI for my own marketing?

One tip would be to integrate AI technologies such as machine learning or NLP (Natural Language Processing) into your marketing strategy. For example, you can use AI tools to better understand your target audience and create personalized content and offers. You can also train AI models to automatically generate content or optimize social media campaigns. Another option would be to use chatbots or virtual assistants to interact with customers. There are many ways you can use AI in your marketing, it's all about identifying your goals and resources and choosing the right tools.

ChatGPT, wow, that was a really interesting conversation. Thank you very much for telling us so much about yourself and your job opportunities. Perhaps you could take a look into the future at the end. How will ChatGPT develop, what else can we expect from you in the future?

It is a pleasure to have helped you. I am sure that the development of AI will continue to progress in the future and that my skills will also continue to develop. In the future, we can expect AI systems like me to be even better at understanding and generating natural language, as well as mastering more complex tasks. There are also likely to be further advances in areas such as computer vision and machine learning, which will further increase the performance of AI systems.

ChatGPT, thank you very much, I look forward to talking to you again in the future!

I also look forward to our next conversation and am always available to help you with your questions.