Olga Fink Professor of Intelligent Maintenance and Operations Systems, EPFL Julien Weissenberg Founder, Deep Tech Experts Generative arti昀椀cial intelligence (AI) is a powerful The newest developments involve autonomous AI type of AI that can create new and original content systems that can make important decisions or take by learning patterns in data, using complex signi昀椀cant actions. For instance, AutoGPT is an algorithms and methods of learning inspired by the autonomous AI application using the GPT-4 language human brain. While generative AI is still currently model. AutoGPT can automatically accomplish focused on producing text, computer programming, a user-identi昀椀ed goal by dividing the goal into 4 images and sound, this technology could be smaller tasks and employing tools like internet applied to a range of purposes, including drug searches or text-to-speech technology. The design, architecture and engineering. growing integration of generative AI technologies, particularly autonomous AI, into multiple aspects For example, at the time of this writing, initial work of people’s daily lives, is generating both public has been published on generating candidate drug excitement and concern. 5 and molecules targeting particular conditions on creating pictures of imaginary buildings or on To build public trust in generative AI, applications generating interior design. NASA engineers are should meet agreed-upon professional and ethical currently working towards AI systems that can standards. Generative AI systems represent the construct lightweight space昀氀ight instruments, data they were trained on and the conventions achieving a 10-fold reduction in development governing society at that time. Care should be time while simultaneously improving structural taken to mitigate AI bias based on training data, performance.6 Generative AI technologies may with a focus on including “outlier” data and novel even impact the food industry and the design of societal conventions. Further, the decision-making everyday objects, from furniture to appliances. In processes of an application should be easy to scienti昀椀c research, generative models could facilitate understand, an application’s goals should be breakthroughs by improving experimental design, clearly disclosed to operators and end users, and identifying relationships between data elements individual privacy should be respected. Ethical and creating new theories. For example, recently guidelines and governance structures must developed AI algorithms can translate a mathematical be developed to mitigate potential harm and formula into plain English or analyse brain activity ensure that technical progress is balanced with data to generate drawings of the objects that human responsible use. Finally, copyright attribution must participants are holding in mind.7,8 be addressed so that proper credit is given to AI designers, creators of training data and authors High school and university students are using of instructions for using the applications. generative AI more frequently, with some institutions forbidding their use while others are integrating With the correct controls in place, generative AI can 9 or generative models into teaching practices provide more time for creativity, demonstrate the even training students to master these tools. Used boundaries of knowledge, and act as a sparring properly, generative AI can create personalized partner to challenge conventional thinking. curricula that adjust to student skills and learning progress while encouraging critical thinking, igniting Image: creativity and harnessing novel ideas. Impact Generative AI can 昀椀ngerprint synthesize what it has learnt In the workplace, the use of AI-based language from gigantic data sets models like the recently popular ChatGPT or its People to create and intuit new successors can increase productivity and improve information in novel ways. Credit: Midjourney and output quality, restructuring human tasks towards Studio Miko. idea generation and editing as opposed to rough Equity Planet 10 drafting. Generative AI technologies speci昀椀cally Prompt (abbreviated): bene昀椀t low-ability workers and can increase job “Glitched image overlaying satisfaction and self-ef昀椀cacy. Given the potential various styles of painting, for productivity gains resulting from adopting these photography and 3D new technologies, it’s crucial to acknowledge the renderings”. likelihood of job displacement. As such, policies Read more: and programmes that support workers in their efforts Discover expert analysis to upskill and reskill are essential in ensuring that the related to generative bene昀椀ts of technological innovation are widely shared Prosperity arti昀椀cial intelligence on Industry the Strategic Intelligence and that workers are equipped with the skills needed Platform. to thrive in the changing job market. Top 10 Emerging Technologies of 2023 11
Top 10 Emerging Technologies of 2023 Page 10 Page 12