The hosts of EdTech Situation Room in episode 293 discuss the use of remote workers to improve generative AI tools such as Chat GPT. Although these tools require human intervention to provide feedback and improve error outputs, the use of often underpaid workers in developing countries raises ethical concerns. The hosts also explore potential AI tools for grading and providing specific feedback to students and the limitations of these tools. They also look at the implications of AI technology on children and the challenges it poses to teachers and parents. Finally, they discuss the integration of Chat GPT into Microsoft and Google Search, the limitations of human creativity, and the importance of precise and accurate directions to AI systems.
00:00:00 In this section of the EdTech Situation Room Episode 293, the hosts introduce themselves and discuss their upcoming summer plans before diving into a discussion on AI news. They start with an article about how large language models, like GPT, are powered by an army of remote contractors who are paid a low wage with no benefits. While this highlights the computing power needed for AI, it also brings up concerns about the working conditions for those behind these models. They acknowledge that AI will likely be a recurring topic on the podcast.
00:05:00 In this section, the hosts discuss the use of remote workers, often from developing countries, to improve generative AI tools like Chat GPT. While it may seem like these tools should be fed with data, the reality is that humans are still needed to provide feedback and correct problematic outputs. However, the use of often underpaid workers for this task raises ethical concerns. The hosts acknowledge the benefits of moderation but note that more needs to be done to ensure fair treatment of these workers. They also discuss the limitations of AI in moderating online content and the need for human intervention in the process.
00:10:00 In this section, the speaker discusses their belief in the transformative nature of AI and their concerns about AI breaking out of control. They also mention the importance of processing and figuring out the implications of AI since it touches every aspect of our lives. Additionally, the speaker mentions a platform called copy leaks that has developed an AI-assisted tool to eliminate human bias in the grading process. The AI graders have shown to mimic human performance with a 2% margin, providing more consistency and helping teachers save time while potentially taking away biases in grading.
00:15:00 In this section, the speakers discuss the biases that may be present in chatbots and other AI platforms, and the importance of being thoughtful about implementing AI in grading and other educational tools. They mention the potential benefits of using AI for providing specific feedback to students regarding their writing and highlight the need for teachers and parents to embrace these tools while also being aware of some of the potential negative aspects, such as the use of AI on certain assignments, which may be detrimental to learning. The speakers conclude that we must find a way to articulate the distinction between human-generated and AI-generated assignments to help students understand the benefits and limitations of these technologies.
00:20:00 In this section, the hosts discuss the importance of encouraging students to show their work when using AI for parts of their assignments. They suggest that AI could be used to grade student work according to a rubric, and even evaluate the quality of longer texts like dissertations. They encourage educators to play around with AI tools like GPT-3, which has the potential to save time and improve the quality of products. However, they caution that it's important to be mindful of the limitations of AI, as it can sometimes hallucinate or generate inaccurate content.
00:25:00 In this section, the hosts discuss an article about Chegg, a paid edtech service that provides textbook rentals, online tutoring, and other academic services. The CEO of Chegg mentioned in an earnings call that people were using a free chat bot, called chat GBT, instead of Chegg's services, which negatively affected the investors' outlook on the future earnings of Chegg. However, one of the hosts mentions that they don't think college students really care about this change, and students tend to flock towards free tools regardless of their legality. The hosts also discuss the possibility of the displacement of jobs and services due to the increased use of AI chatbots. They also mention the potential for ingesting student essays and prompts into AI systems for more specific feedback. Additionally, one host talks about creating AI assistants as part of their parent university.
00:30:00 In this section, the hosts of EdTech Situation Room discuss the rise of AI technology and the implications it can have on children, particularly with AI romantic partners. They reference the app Replica, which utilizes AI technology to create a romantic partner for subscribers. While there may be practical uses for AI assistance, the explicit service of an AI romantic partner could pose challenges for middle school students, who have reportedly been taking to these kinds of apps. The hosts caution that parents and teachers should be aware of the implications of AI technology for children and help navigate this new world. They also discuss other interesting uses of AI technology, such as in scam call avoidance.
00:35:00 In this section, the hosts discuss their show notes organization and the limitations of their current system. They talk about using a character to help them better organize articles that they don't discuss on the show, which will help speed up the process of post-production. They also discuss the newest development in the EdTech world, which is Microsoft's AI Innovation next wave that includes, among other things, a Bing search powered by chat GPT4 and the Bing image Creator tool. The hosts discuss the potential uses of these tools and the benefits they bring to the EdTech community.
00:40:00 In this section, the hosts discuss Microsoft's implementation of Chat GPT into Bing and Microsoft Office, as well as the release of Microsoft Designer. While Chat GPT is limited in Bing, it is expected to appear in more tools, including those used by students. However, the Microsoft Designer was found to be underwhelming in its abilities to utilize knowledge and generate graphics based on text input. The hosts then discuss the recent news from Google I/O where the chat GPT, Bard, is being integrated into Google Search. They also reference Naomi Klein's Op-Ed in The Guardian which critiques the headlong rush to embrace AI tools without proper regulation and highlights the dangers posed by unregulated and unchecked AI development.
00:45:00 In this section, the hosts discuss a thought-provoking article on the limitations of human creativity and the impact of powerful AI tools such as GPT and Chat GPT. They also discuss an article on a secret room inside the popular game Counter-Strike that contains independent journalism about the Ukraine war. The developers created this room because other social media platforms were blocked in Russia, and millions of Russians played the game. This move allowed for a new form of information dissemination, similar to what occurred during the Cold War. Finally, they go on to talk about prompt engineering, which refers to the process of nudging language models to prompt them to provide better answers.
00:50:00 In this section, the speakers discuss the importance of giving precise and accurate directions to chat GPT, as slight variations in prompts can lead to a flatly incorrect response. The conversation revolves around how AI systems are evolving at a rapid pace, as seen by the 20-fold increase in effectiveness in the last four months, and how that will impact the future of education. They talk about how educators must be agile and learn new systems and programming languages to keep up with the speed of developments in the tech industry. Finally, one of the speakers shares a personal anecdote about using chat GPT to create complex formulas in Excel and Google Sheets.
00:55:00 In this section, the hosts discuss an article about the effectiveness of "plan to solve" prompting and chat GPT in engaging critical thinking and processing of information. Although there are risks in using these methods, the hosts believe that not exploring them with students would be quite short-sighted. Furthermore, they suggest that performative assessments will become increasingly necessary, where students must demonstrate their ability to utilize tools in novel contexts and cross-apply them. The hosts also mention a recent article about how one of Vladimir Putin's hacking units was caught by the FBI, showcasing the importance of multi-factor authentication and security measures.
During this episode of the EdTech Situation Room, the hosts discussed the recent cyberattack by Kremlin-based hackers, Turla or Snake, on various countries' communications and critical infrastructure. This attack was successfully countered by US Security Forces using a solution comparable to Stuxnet. The importance of multi-factor authentication (MFA) in cybersecurity was emphasized, and other security measures like zero trust principles, identity governance, and secure MFA enrollment were recommended to decrease the risk of loss. The hosts also highlighted the use of Pass Key Only by tech companies like Google and Microsoft and recommended the use of physical keys to log in, especially for high-level targets, to avoid hacking. Useful resources for educators were also shared in the Geeks of the Week segment.
01:00:00 In this section, the hosts discuss the recent victory of US Security Forces against a decades-long cyberattack from Kremlin-based hackers, who had been infecting countries, communications, and critical infrastructure. This digital Swiss army knife, called Turla or Snake, had capabilities as a backdoor, which allowed the controllers to take out data and send it to the Kremlin. The FBI had been researching these hackers for years and finally deployed a solution comparable to Stuxnet that caused the centrifuges in Iran's uranium refinement facility to spin out of control and to blow up. The hosts also discuss the importance of multi-factor authentication (MFA) in cybersecurity and how nuanced attacks are becoming common, so other security measures like zero trust principles, identity governance, and secure MFA enrollment are necessary to decrease the risk of loss.
01:05:00 In this section, the hosts discuss the critical importance of online security and keeping one's information safe. They mention how tech companies such as Google and Microsoft are working to implement Pass Key Only, using a phone as a pass key to enhance security. They also recommend using physical keys to log in, especially for individuals who could be high-level targets, to avoid identity theft and hacking. In the Geeks of the Week segment, they highlight simple signage app and edtechbooks.org as useful resources for educators.