Summary of Title: This Week in AI - 28 June 2024

00:00:00 - 00:35:00

In the "This Week in AI" video from June 28, 2024, hosts Steve Hargadon and Reed Hepler discuss the latest developments and implications of AI in various industries, including education, music, video creation, and expertise. They reflect on the disruption caused by AI-generated content, the potential impact on grassroots creators, and the blurring lines between reality and virtual representation. The conversation covers the recent acquisition of Multi by OpenAI , the advancements in AI technology, and its potential impact on communication, productivity, and expertise. The hosts also touch on the concerns about the diminishment of expertise and the role of humans in interpreting and applying AI-generated information. They end the discussion by acknowledging the rapid evolution of the relationship between humans and AI and its significant implications for the future of work and society.

• 00:00:00 In this section, Steve Hargadon and Reed Hepler discuss the recent "Teaching and Learning with AI" conference they organized, which had over 9,700 sign-ups. They also share their thoughts on creating shorthand phrases for using generative AI tools, such as "LOM" (pronounced "lambed") and "Jen." The conversation then shifts to the ongoing legal dispute between record companies and AI music generators, Suno and Unio, over copyright infringement. The hosts discuss the potential implications of the case and its comparison to the ride-sharing industry's fight against new business models.
• 00:05:00 In this section of the "This Week in AI" video from 28 June 2024, Steve Hargadon and Reed Hepler discuss the impact of AI on various creative industries, including music and visual media. They reflect on the potential disruption caused by AI-generated content, particularly in terms of artistic expression, revenue, and cultural implications. Hargadon shares his personal experience with the Wayne Johnson trio and expresses his willingness to consume AI-generated music if it could replicate the experience of listening to his favorite band. They also discuss the recent AI-generated video by Toys R Us, which received mixed reactions but showcased the potential for completely AI-generated storylines. Hepler expresses concerns about the potential elitism surrounding AI-generated content and the impact on grassroots creators. Despite these concerns, they both agree that AI opens up enormous opportunities for creative people to produce compelling content with fewer financial barriers.
• 00:10:00 In this section of the "This Week in AI" video from June 28, 2024, Reed Hepler and Steve Hargadon discuss the future of video creation and the role of AI in shaping the output of AI models. Hepler predicts that video creation will become more commonplace, with many people trained to create videos using AI. He also mentions OpenAI's use of reinforcement learning with human feedback to fine-tune AI models and make their output more coherent and accurate. Hargadon expresses concerns about the potential influence of large language models and the possibility of certain industries or governments shaping the answers provided by these models. Hepler adds that the standard training process for AI tools, including OpenAI's RLHF, raises questions about the specific purposes and conditioning of these models. The conversation ends with a hope that open-source versions of AI will allow people to examine the reasoning behind the models' answers and ask questions without being warned off potential dangers.
• 00:15:00 In this section, Steve Hargadon and Reed Hepler discuss the recent acquisition of a small company called Multi by OpenAI, which specializes in a program that allows AI to share and control a desktop. The hosts pondered the reason behind the acquisition until they realized that OpenAI aims to transform ChatGPT into an AI agent that can influence a user's computer, making it a collaborator. This capability could potentially displace other AI companies like Synestia, which offers an avatar that attends meetings and takes notes in the user's place. Synestia also has more features, such as converting a user's knowledge base into videos and creating personal AI avatars with hands. The idea of generating a likeness of oneself through AI raises questions about the blurring lines between reality and virtual representation, reminiscent of science fiction films like "Surrogates."
• 00:20:00 In this section of the "This Week in AI" video from June 28, 2024, Reed Hepler and Steve Hargadon discuss the advancements in AI technology and its potential impact on communication and productivity. They reflect on how AI has evolved from simple video conferencing avatars to more sophisticated tools capable of understanding and responding to objectives. Hepler shares his recent blog post about the importance of conversation steering rather than prompt engineering in interacting with AI. Hargadon agrees, adding that the era of prompt engineering may soon diminish as AI becomes more integrated into various programs and better understands users' needs. They also mention the increasing productivity gains from AI usage, as evidenced by a recent article in ZDNet.
• 00:25:00 In this section of "This Week in AI - 28 June 2024", Steve Hargadon and Reed Hepler discuss the productivity gains and job enjoyment reported by workers using AI, as well as the potential for AI to help adhere to policies and procedures. They also touch on the idea that AI could improve consistency and even boost creativity, while acknowledging potential concerns about creativity stifling. Steve Hargadon shares an anecdote about using Chatgpt for a health concern instead of consulting an expert, raising questions about the future of expertise and motivation for in-depth content creation when AI can absorb and provide answers. The conversation delves into the differences between AI's access to information with no memory loss and a human's ability to read and write, sparking ongoing discussions about the implications of AI on various aspects of productivity, creativity, and expertise.
• 00:30:00 In this section of the "This Week in AI" video from June 28, 2024, Steve Hargadon and Reed Hepler discuss the implications of advanced AI on expertise and knowledge sharing. Hargadon expresses concerns about the potential diminishment of expertise when AI can absorb and produce content without referencing original sources. Hepler argues that experts' roles may shift from content creators to curators and amalgamators of information. They also discuss how AI can outperform humans in certain tasks, such as coding advice or content creation, potentially leading to the "death of expertise." However, they acknowledge that human expertise and perspective will still be valuable in interpreting and applying AI-generated information, especially in fields like medicine. The conversation ends with a recognition that the relationship between humans and AI is evolving rapidly and that the implications are significant for the future of work and society.
• 00:35:00 In this section of the "This Week in AI" video from June 28, 2024, Steve Hargadon and Reed Hepler discuss the potential impact of AI on education and who will control the outcome. Hargadon raises concerns about the influence of commercial and political interests on the development and use of AI, particularly in the context of Google's new learning tool, Google LM. Hepler questions how AI will change civilization and suggests that it may become a significant source of knowledge and education. The conversation also touches on the paradox of education and the role of AI in empowering individuals and filling the "pale model" of top-down information delivery. The speakers also mention upcoming events, including a professional productivity boot camp with over 30 AI tutorials.