Wolfgang Zehrt
The Di5ru.pt blog

In preparation for an in-depth report on AI and the media, we spoke to several experts in the field via email. One such expert is Wolfgang Zehrt, AI and Content Expert in Germany.

What are your current most common use cases of AI? 

We currently work for five publishing houses and a total of 40 publications: From a regional daily newspaper to a leading German national newspaper to newspapers and trade portals in the USA and Asia. 

We publish about 1,000 articles daily. Of these, about three-quarters are articles from data generated together with graphics and the future with AI-generated images. 

About a quarter of the articles are existing text revised by our AI tool. These are mainly press releases formulated into factual reports or announcements of court decisions transferred into everyday language. 

In addition, we have more and more improvements in translations that go through our program. It does not matter what the input language and the output language are. In our sandbox, www.digitaldaily.de, all kinds of automated content are visible. This includes text done by a prototype of GPT-4.

For the media, it is crucial to remember that OpenAI’s models and tools do not work with current data because they are not real-time systems, they are trained on a large dataset with a certain knowledge cut-off. 

Additionally, the models are not updated in real-time, so they may not reflect the most recent changes in the data they were trained on. To ensure that the models are working with the most current data, the data should be updated and the models should be retrained. This process can be time-consuming and requires significant computational resources. 

It’s also important to note that OpenAI’s models, like any other AI model, can’t account for all the events, trends, or information that is happening in the current moment, as they are not able to incorporate new knowledge or experience, they only can provide information based on the data they have been trained with. For this reason, by using a self-developed tool we integrate our real-time data into OpenAI systems. The whole process is still 100% automated.

How else will you use AI in the coming year? 

AI is expected to be increasingly used in fields such as healthcare, finance, and transportation. Additionally, AI is expected to play a greater role in automating repetitive tasks and decision-making processes. 

OpenAI’s models and tools have the potential to automate a wide range of tasks and processes, which could lead to some jobs being replaced. 

Some examples of jobs that may be impacted by OpenAI include:

  1. Data entry: OpenAI’s models can be used to automatically process and analyze large amounts of data, which could lead to the automation of data entry jobs.
  2. Content creation: OpenAI’s language model GPT-3 can be used to generate human-like text, which could lead to the automation of writing jobs such as journalism, content writing, script writing, and more.
  3. Translation: OpenAI’s models can be used to translate text between different languages, which could lead to the automation of translation jobs.
  4. Customer service: OpenAI can be used to create chatbots and automated responses for customer service, which could lead to the automation of customer service jobs.
  5. Research: OpenAI’s models can be used to process and analyze large amounts of data, which could lead to the automation of research jobs.
  6. Data analysis: OpenAI can be used to make predictions, identify patterns and uncover insights that could lead to the automation of data analysis jobs.

What excites you most about the longer-term future of AI, and why?

I am excited about the potential for AI to improve many aspects of our lives by making them more efficient and convenient. The ability of AI to process and analyze large amounts of data has the potential to uncover new insights and knowledge that would be impossible for humans to discover alone. 

OpenAI is exciting because it is a cutting-edge research organization that is working to develop new AI models and technologies that have the potential to improve many aspects of our lives. Some reasons why I find OpenAI exciting:

  1. Advancements in natural language processing: OpenAI’s language model, GPT-3, has shown significant improvements in natural language processing, which has the potential to improve human-computer interactions and make it easier for people to access and understand information.
  2. Advancements in machine learning: OpenAI is working on developing new machine learning models that are more efficient, accurate, and scalable, which has the potential to improve a wide range of applications such as speech recognition, computer vision, and more.
  3. Advancements in AI safety: OpenAI is also working on developing new methods for ensuring the safety of AI systems and the alignment of their goals with human values, which will prevent potential negative consequences of AI.
  4. Advancements in AI democratization: OpenAI is making its models and tools available to a wide range of developers and researchers, which is helping to democratize access to AI and make it more widely available to people and organizations.

What concerns you most about the longer-term future of AI, and why?

One concern about the longer-term future of AI is the potential for it to be used in ways that are harmful or unethical. Additionally, there are concerns about the potential for AI to automate many jobs, leading to increased unemployment. 

Some reasons why one might be concerned about GPT-3 f.e. include:

  1.   Misinformation: GPT-3 can generate human-like text, which can be used to create convincing fake news, misinformation, and propaganda.
  2.   Lack of accountability: GPT-3 generates text without any inherent bias or purpose, so it can be used to produce content that is harmful or offensive. It is important that creators and users of GPT-3 must take responsibility for the text it generates and not use it for malicious purposes.
  3.   Job displacement: GPT-3 has the potential to automate some jobs that involve writing or content creation, which could lead to job displacement.
  4.   Privacy concerns: GPT-3 requires large amounts of data to train its models, so there are concerns about the privacy of the data that is used to train it.
  5.   Ethical concerns: GPT-3 is a powerful tool that can be used to generate text that is very similar to human writing, which raises ethical concerns about the authenticity of the text it generates and the implications of using it.

 What are some of the best AI tools (already available for use) that you can recommend to media companies to investigate immediately?

Some of the best AI tools that I can recommend to media companies to investigate immediately include:

  • IBM Watson Studio: an AI platform that allows companies to build, train, and deploy machine learning models.
  • Google Cloud AI: offers a wide range of AI services such as natural language processing, image and speech recognition, and machine learning.
  • Microsoft Azure: an AI platform that provides tools and services for building, deploying, and managing AI models.
  • OpenAI GPT-3: a language model that can be used to generate human-like text and can be integrated into a wide range of applications.

Try these tools for:

  1. News generation: OpenAI’s GPT-3 can be used to generate news articles, summaries, and even videos, allowing media companies to create news faster and more efficiently.
  2. Social media: OpenAI can be used to create chatbots and automated responses for social media platforms, making it easier for media companies to interact with their audience.
  3. Advertising: OpenAI can be used to generate creative and engaging content for advertisements, increasing their effectiveness.
  4. Search engine optimization (SEO): OpenAI can be used to generate keywords and content that are optimized for search engines, helping media companies to improve their visibility on the web.
  5. Voice assistants: OpenAI can be used to create voice assistants that can understand natural language and respond appropriately, making it easier for media companies to create interactive experiences for their audience.
  6. Video captioning: OpenAI can be used to automatically generate captions for videos, making it easier for media companies to create accessible content.

There is currently a big discussion about how dangerous OpenAI could become for Google. Anyone who knows that real-time data and real-time information can currently only be integrated into ChatGPT responses with considerable technical effort knows that this will not be the case for many years. 

From OpenAI plus self-developed solutions, a “Google” is feasible, but only for limited subject areas. We implement this for many clients and are still far from being able to replace Google :)