14 Albert Einstein once said: “We cannot claim to fully understand a concept unless we are able to explain it.” The AI technology can be used not only to create new content, but also to help review and edit existing text. For example, a person composing an email can use AI to automatically rewrite the message in a more professional style. The type of text that can be analyzed and generated is not limited to human language as it can also include other types of languages, such as computer codes. This is all thanks to a specific type of AI, “Generative AI”, so called for its ability to generate clear and well organized responses by making probabilistic guesses based on billions of examples from digital sources. Not only can it write sentences, but it can write essays on any topic, compose poems, summarize scientific papers, share jokes, and even debate ethical dilemmas. It can also answer follow-up questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests. As this new technology can be easily deployed using widely used platforms such as PCs and smartphones, it will likely be highly disruptive. The business community has already identified countless opportunities to deploy Generative AI in support of business processes such as customer support, search tools, document creation and analysis, copywriting, media production, and coding. Images and videos can also be generated by editing existing images and videos based on the user’s requests. The AI software model is trained by submitting millions of digital images with captions that describe them, with the goal of mapping links between the images and the words. It then generates images based on the directions provided by the user. Similar to what graphic designers can do with tools such as Photoshop, the tool can instantly create new content based on photos, videos, illustrations, artwork, etc. This technology can also be used to create videos of virtual people that are absolutely realistic. Such videos can be created ad-hoc for presentations, advertisements, movies, and TV shows, with countless customization options to better serve a specific audience. For example, a promotional video for a product can be produced in seconds with highly realistic AI-generated human-like presenters and deployed across the world with automatic and real-time customizations based on the audience’s preferences, such as skin color, clothing, language, etc. Sounds such as music and even human voices can be created by AI models as well. AI can easily compose and play original music pieces with new melodies, styles, and instrument sounds never heard before. Highly realistic human voices can be generated by AI reproducing the voice of a specific person or creating new timbres with the option to customize gender, accent, etc. They can then be used in applications such as multilingual advertisements, real-time automated speech dubbing, etc. The artistic value of the content generated by AI models is often so high, that many have started wondering about the meaning and nature of creativity itself. If AI can produce something new and with significant artistic value, how can it not be considered a form of creativity? More in general, it is clear that tasks intrinsically tied to human culture, such as the ability to write a story or paint a picture, will no longer be an exclusive prerogative of humans. Will AI end up generating not just content, but also culture? Despite its impressive features, Generative AI still lacks the ability to think, to understand how the world works, or any of the meaning of the material being generated. While the work products of Generative AI models are often indistinguishable from human-generated content, Generative AI lacks any kind of common sense or humanlike agency. AI is all syntax (arranging the words in the right order) and no semantics (understanding of the meaning). While the AI model is incredibly good at predicting the correct sequence of words in a sentence, it has zero knowledge of what the words and the sentences mean. It is just a guessing machine. There is a voice but not a brain able to comprehend the meaning of what is being said. Any textgenerating AI needs to be treated more like a powerful game than as a dependable source of information, so human critical thinking abilities will be needed more than ever to fact check and correct all the false or misleading information generated by AI. Similar types of reliability issues exist for AI-generated images, as well. It is not uncommon, for example, for AI-generated portraits to have evident errors, especially for details such as the shape of people’s hands. The model does not have any understanding of what human hands actually are and it is not able to identify mistakes. The automation of quality control activities to identify and correct idiotic or absurd content is not possible with current technology and it will likely take a major advancement in AI technology to get us there. With the delegation of a large component of the creative process to AI, there will be a significant impact on those whose primary job is to write articles, draft legal documents, create visuals, compose music, perform on screen (film, television, etc.), and even develop computer code. For example, the time needed to research, compose, and illustrate a news piece can now be easily scaled down from days to just seconds. Even if the quality of the work product is too low to be published and it requires time to make sure it is ready, the time advantage is still very attractive. Content generating technologies make it easier for bad actors to create and widely distribute offensive or inflammatory content, misinformation, spam, cyber threats, clickbait campaigns, scams, and deep fakes. Generative AI is excellent at generating images, including offensive ones. For example, Generative AI allows you to instantly edit phoWhat to Expect from ArtificiaI Intelligence: Present and Future October 2023 Tech Talks The artistic value of the content generated by AI models is often so high, that many have started wondering about the meaning and nature of creativity. (Continued on Page 16) By Matteo Di Michele *
RkJQdWJsaXNoZXIy MTI4OTA5