October 20, 2025 • 7 mins
This week in AI: OpenAI's Sora 2 ignites a firestorm in Hollywood as studios and unions raise alarm over AI-generated videos featuring real people's likenesses. Google unveils Gemini 2.5 Computer Use, a breakthrough model that can actually click, scroll, and navigate websites like a human, no APIs required. eBay launches a £3 million initiative giving 10,000 small businesses free access to ChatGPT Enterprise, leveling the playing field for sellers who want to harness AI.
Plus, OpenAI rolls out ChatGPT Search to all users (including free tier) and introduces Advanced Voice Mode with real-time video input, bringing universal accessibility features that let you point your phone at anything and get instant voice responses.
From copyright battles to browser automation breakthroughs, this week showcases both AI's incredible potential and the tensions it creates. Your complete AI news roundup in under 10 minutes.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Hi everyone.
(00:00):
I'm Andy and this is the AI Breakdown.
.233688478Welcome to your weekly news edition where I'll cover what happened in AI last week, why it matters all in less than 10 minutes open. 4 00:00:14,844.233688478 --> 00:00:24,924.233688478 AI has stirred up Hollywood with the release of SO two, a groundbreaking text to video tool capable of inserting real people and characters into AI generated videos. 5 00:00:25,464.233688478 --> 00:00:32,844.233688478 It's received a rather icy reception from Hollywood Studios and unions who are up in arms about copyright and privacy concerns. 6 00:00:33,264.233688478 --> 00:00:40,524.233688478 The fear is all about unapproved use of protected likenesses, which rightly so has sparked significant backlash. 7 00:00:40,854.233688478 --> 00:00:43,974.233688478 The implications of SOA to extend beyond just Hollywood. 8 00:00:44,334.233688478 --> 00:00:49,674.233688478 This development could redefine IP controls and transformation in creative processes. 9 00:00:50,124.233688478 --> 00:00:57,534.233688478 SOA two's capabilities are remarkable, generating realistic video content that previously required hefty budgets and large teams. 10 00:00:57,984.233688478 --> 00:00:59,629.233688478 Yet it's a double-edged sword. 11 00:01:00,429.233688478 --> 00:01:06,249.233688478 If not handled carefully, it might blur the lines of intellectual property and creative ownership Even more. 12 00:01:06,969.233688478 --> 00:01:17,529.233688478 Charles Rifkin, chairman of the Motion Picture Association, made his sentiments quite clear when he said OpenAI needs to take immediate and decisive action to address this issue. 13 00:01:18,69.233688478 --> 00:01:27,789.233688478 The saga highlights the constant tension between advancing technology and existing legal and ethical frameworks in creativity, and it's one to watch closely as it develops. 14 00:01:29,418.425014448 --> 00:01:33,78.425014448 Google has just rolled out a new AI model called Gemini 2.5 15 00:01:33,78.425014448 --> 00:01:38,868.425014448 computer use an innovative AI model that enables bots to interact with web pages like a human. 16 00:01:39,288.425014448 --> 00:01:40,578.42501445 It's not your regular bot. 17 00:01:40,938.42501445 --> 00:01:47,298.42501445 It's designed to let AI agents navigate websites and apps by actually clicking, typing, and scrolling in a browser. 18 00:01:48,168.42501445 --> 00:01:51,348.42501445 It's a big leap forward in AI driven process automation. 19 00:01:52,8.42501445 --> 00:02:03,438.42501445 Unlike typical bots that rely on APIs, this model visually interprets a webpage like a human would, looking at the screen and can perform tasks such as filling out forms and submitting information. 20 00:02:04,128.42501445 --> 00:02:06,468.42501445 Google demonstrated Gemini 2.5 21 00:02:06,468.42501445 --> 00:02:13,968.42501445 performing actions like playing a web game and browsing news autonomously, using a loop of reading the screen and taking actions. 22 00:02:14,658.42501445 --> 00:02:18,18.42501445 The model is available in preview via Google's AI developer tools. 23 00:02:18,723.42501445 --> 00:02:25,353.42501445 Vertex AI and AI studio enabling developers to build agents that carry out multi-step tasks on websites. 24 00:02:26,13.42501445 --> 00:02:28,503.42501445 Developers are buzzing with excitement about the potential. 25 00:02:28,503.42501445 --> 00:02:28,893.42501445 Here. 26 00:02:29,253.42501445 --> 00:02:34,263.42501445 It cuts out the meticulous task of integrating specific APIs for every function. 27 00:02:34,623.42501445 --> 00:02:36,633.42501445 Make an app development more efficient. 28 00:02:37,293.42501445 --> 00:02:47,133.42501445 Imagine an AI gathering competitor pricing from multiple websites, updating your knowledge base and generating battle cards without needing API access to any of those sites. 29 00:02:47,733.42501445 --> 00:02:54,423.42501445 Or an AI automatically pulling receipts from your email, logging into your expense system and filing your monthly reports. 30 00:02:55,23.42501445 --> 00:02:58,173.42501445 Now I've been trying to crack this type of problem myself for months. 31 00:02:58,503.42501445 --> 00:03:07,113.42501445 I've experimented with chat GPT Agent Mode, wrestled with Selenium browser automation, tried various workarounds, all with pretty mixed results. 32 00:03:07,443.42501445 --> 00:03:08,943.42501445 Sometimes it works brilliantly. 33 00:03:08,973.42501445 --> 00:03:10,863.42501445 Sometimes it's a complete disaster. 34 00:03:11,403.42501445 --> 00:03:13,743.42501445 So I'm excited to try out this Gemini update. 35 00:03:14,193.42501445 --> 00:03:16,38.42501445 This technology is a major step forward. 36 00:03:16,743.42501445 --> 00:03:19,383.42501445 Pushing the boundaries of what digital assistance can do. 37 00:03:20,809.46007452 --> 00:03:27,409.46007452 eBay's recent initiative AI Activate is a bold move towards leveling the playing field for small businesses. 38 00:03:27,919.46007452 --> 00:03:32,629.46007452 The 3 million pound program launched in the UK will onboard up to 10,000. 39 00:03:32,629.46007452 --> 00:03:41,539.46007452 eBay sellers, giving them free 12 month access to chat GPT Enterprise and Custom AI productivity tools developed by eBay. 40 00:03:42,454.46007452 --> 00:03:56,494.46007452 Participants will also receive training both virtual sessions and later in person workshops on how to use AI for tasks like financial analysis, writing, marketing, copy, researching market trends, and improving product listings. 41 00:03:57,124.46007452 --> 00:04:06,844.46007452 Notably, eBay is the first major marketplace to provide open AI's enterprise grade tools at no cost to its sellers aiming to boost their efficiency and growth online. 42 00:04:07,969.46007452 --> 00:04:18,79.46007452 The program was crafted after surveying small businesses, which indicated strong interest in ai, but uncertainty about how to implement it, and it builds on eBay's own use of ai. 43 00:04:18,439.46007452 --> 00:04:27,529.46007452 Over 10 million sellers have already used eBay's generative AI features to create more than 300 million product listings, contributing billions in sales. 44 00:04:27,919.46007452 --> 00:04:30,439.46007452 This initiative is more than just an act of generosity. 45 00:04:30,769.46007452 --> 00:04:33,199.46007452 It's a strategic move that could significantly. 46 00:04:33,559.46007452 --> 00:04:39,349.46007452 Boost productivity in areas like marketing and market research for small to medium enterprises. 47 00:04:39,829.46007452 --> 00:04:47,59.46007452 By doing so, it opens up the world of artificial intelligence to businesses that might not have had the resources to engage with such technology. 48 00:04:47,59.46007452 --> 00:04:54,19.46007452 Otherwise, the program has been met with positive feedback from small businesses being hailed as a potential game changer. 49 00:04:54,469.46007452 --> 00:04:58,39.46007452 Such moves are likely to encourage other platforms to offer similar tools. 50 00:04:58,564.46007452 --> 00:05:12,664.46007452 This kind of reminds me of what's happening in the US public sector where tech giants are offering AI products to government agencies for free or nearly free, essentially trying to get themselves ingrained early and ultimately forced dependence on one supplier. 51 00:05:13,84.46007452 --> 00:05:14,884.46007452 What E eBay is doing here is clever. 52 00:05:15,184.46007452 --> 00:05:19,534.46007452 They're democratizing AI access to keep their platform competitive and diverse. 53 00:05:19,924.46007452 --> 00:05:21,929.46007452 Small sellers get enterprise level tools. 54 00:05:22,684.46007452 --> 00:05:29,344.46007452 eBay keeps its marketplace vibrant and they avoid becoming just another platform dominated by a few AI powered Giants. 55 00:05:30,858.60137506 --> 00:05:36,48.60137506 OpenAI has released new updates for chat GPT, making it more powerful and versatile than ever. 56 00:05:36,798.60137506 --> 00:05:46,8.60137506 It announced that the AI's built-in web browsing feature known as Chat GPT Search is now available to all users, including free tier globally. 57 00:05:46,848.60137506 --> 00:05:53,28.60137506 This means anyone can use chat GPT to get fast, timely answers with relevant web sources on current events. 58 00:05:53,373.60137506 --> 00:05:59,253.60137506 Information or anything on the internet effectively turning chat GPT into a direct Google competitor. 59 00:06:00,3.60137506 --> 00:06:06,543.60137506 OpenAI even released a browser extension so users can set chat GPT as their default search engine in Chrome. 60 00:06:07,233.60137506 --> 00:06:20,403.60137506 At the same time, OpenAI introduced major multimodal features, including the advanced voice mode, which extends their voice conversations feature, upgrading it to support real time image and video input, and even screen sharing. 61 00:06:21,138.60137506 --> 00:06:28,278.60137506 In practice, a user could point their phone camera or share their screen with chat GPT and ask questions about what they're seeing. 62 00:06:28,698.60137506 --> 00:06:36,378.60137506 For instance, debugging an error on screen of solving a handwritten math problem and chat GPT will respond with voice and visual cues. 63 00:06:36,798.60137506 --> 00:06:38,478.60137506 The timing here is perfect. 64 00:06:38,748.60137506 --> 00:06:46,368.60137506 I literally just released a podcast over the weekend about how blind comedian Chris McCasland is using AI to transform his independence. 65 00:06:47,163.60137506 --> 00:06:58,983.60137506 I looked at how he uses AI vision tools as essential daily assistance, pointing his phone at audiences to gauge their reactions, have an AI describe what's around him, these new chat GPT features. 66 00:06:58,983.60137506 --> 00:07:02,973.60137506 Basically take that specialist accessibility tick and make it universal. 67 00:07:03,483.60137506 --> 00:07:08,433.60137506 Now anyone can point their phone and ask, what am I looking at? And get instant voice responses. 68 00:07:08,943.60137506 --> 00:07:13,533.60137506 As Chris always says, the best accessibility innovations and are benefiting everyone. 69 00:07:13,983.60137506 --> 00:07:16,53.60137506 These updates rolled out within the past week. 70 00:07:16,383.60137506 --> 00:07:26,583.60137506 Or a continuation of chat GPT transitioning from a text only assistant to something like a full service AI helper that can browse, speak, listen, and see. 71 00:07:28,3.00827676 --> 00:07:29,893.00827676 That's all for this week's AI roundup. 72 00:07:30,223.00827676 --> 00:07:34,33.00827676 If you found value in this breakdown, please leave a rating and hit subscribe. 73 00:07:34,33.00827676 --> 00:07:35,23.00827676 See you next week.
Hi everyone.
(00:00):
I'm Andy and this is the AI Breakdown.
.233688478Welcome to your weekly news edition where I'll cover what happened in AI last week, why it matters all in less than 10 minutes open. 4 00:00:14,844.233688478 --> 00:00:24,924.233688478 AI has stirred up Hollywood with the release of SO two, a groundbreaking text to video tool capable of inserting real people and characters into AI generated videos. 5 00:00:25,464.233688478 --> 00:00:32,844.233688478 It's received a rather icy reception from Hollywood Studios and unions who are up in arms about copyright and privacy concerns. 6 00:00:33,264.233688478 --> 00:00:40,524.233688478 The fear is all about unapproved use of protected likenesses, which rightly so has sparked significant backlash. 7 00:00:40,854.233688478 --> 00:00:43,974.233688478 The implications of SOA to extend beyond just Hollywood. 8 00:00:44,334.233688478 --> 00:00:49,674.233688478 This development could redefine IP controls and transformation in creative processes. 9 00:00:50,124.233688478 --> 00:00:57,534.233688478 SOA two's capabilities are remarkable, generating realistic video content that previously required hefty budgets and large teams. 10 00:00:57,984.233688478 --> 00:00:59,629.233688478 Yet it's a double-edged sword. 11 00:01:00,429.233688478 --> 00:01:06,249.233688478 If not handled carefully, it might blur the lines of intellectual property and creative ownership Even more. 12 00:01:06,969.233688478 --> 00:01:17,529.233688478 Charles Rifkin, chairman of the Motion Picture Association, made his sentiments quite clear when he said OpenAI needs to take immediate and decisive action to address this issue. 13 00:01:18,69.233688478 --> 00:01:27,789.233688478 The saga highlights the constant tension between advancing technology and existing legal and ethical frameworks in creativity, and it's one to watch closely as it develops. 14 00:01:29,418.425014448 --> 00:01:33,78.425014448 Google has just rolled out a new AI model called Gemini 2.5 15 00:01:33,78.425014448 --> 00:01:38,868.425014448 computer use an innovative AI model that enables bots to interact with web pages like a human. 16 00:01:39,288.425014448 --> 00:01:40,578.42501445 It's not your regular bot. 17 00:01:40,938.42501445 --> 00:01:47,298.42501445 It's designed to let AI agents navigate websites and apps by actually clicking, typing, and scrolling in a browser. 18 00:01:48,168.42501445 --> 00:01:51,348.42501445 It's a big leap forward in AI driven process automation. 19 00:01:52,8.42501445 --> 00:02:03,438.42501445 Unlike typical bots that rely on APIs, this model visually interprets a webpage like a human would, looking at the screen and can perform tasks such as filling out forms and submitting information. 20 00:02:04,128.42501445 --> 00:02:06,468.42501445 Google demonstrated Gemini 2.5 21 00:02:06,468.42501445 --> 00:02:13,968.42501445 performing actions like playing a web game and browsing news autonomously, using a loop of reading the screen and taking actions. 22 00:02:14,658.42501445 --> 00:02:18,18.42501445 The model is available in preview via Google's AI developer tools. 23 00:02:18,723.42501445 --> 00:02:25,353.42501445 Vertex AI and AI studio enabling developers to build agents that carry out multi-step tasks on websites. 24 00:02:26,13.42501445 --> 00:02:28,503.42501445 Developers are buzzing with excitement about the potential. 25 00:02:28,503.42501445 --> 00:02:28,893.42501445 Here. 26 00:02:29,253.42501445 --> 00:02:34,263.42501445 It cuts out the meticulous task of integrating specific APIs for every function. 27 00:02:34,623.42501445 --> 00:02:36,633.42501445 Make an app development more efficient. 28 00:02:37,293.42501445 --> 00:02:47,133.42501445 Imagine an AI gathering competitor pricing from multiple websites, updating your knowledge base and generating battle cards without needing API access to any of those sites. 29 00:02:47,733.42501445 --> 00:02:54,423.42501445 Or an AI automatically pulling receipts from your email, logging into your expense system and filing your monthly reports. 30 00:02:55,23.42501445 --> 00:02:58,173.42501445 Now I've been trying to crack this type of problem myself for months. 31 00:02:58,503.42501445 --> 00:03:07,113.42501445 I've experimented with chat GPT Agent Mode, wrestled with Selenium browser automation, tried various workarounds, all with pretty mixed results. 32 00:03:07,443.42501445 --> 00:03:08,943.42501445 Sometimes it works brilliantly. 33 00:03:08,973.42501445 --> 00:03:10,863.42501445 Sometimes it's a complete disaster. 34 00:03:11,403.42501445 --> 00:03:13,743.42501445 So I'm excited to try out this Gemini update. 35 00:03:14,193.42501445 --> 00:03:16,38.42501445 This technology is a major step forward. 36 00:03:16,743.42501445 --> 00:03:19,383.42501445 Pushing the boundaries of what digital assistance can do. 37 00:03:20,809.46007452 --> 00:03:27,409.46007452 eBay's recent initiative AI Activate is a bold move towards leveling the playing field for small businesses. 38 00:03:27,919.46007452 --> 00:03:32,629.46007452 The 3 million pound program launched in the UK will onboard up to 10,000. 39 00:03:32,629.46007452 --> 00:03:41,539.46007452 eBay sellers, giving them free 12 month access to chat GPT Enterprise and Custom AI productivity tools developed by eBay. 40 00:03:42,454.46007452 --> 00:03:56,494.46007452 Participants will also receive training both virtual sessions and later in person workshops on how to use AI for tasks like financial analysis, writing, marketing, copy, researching market trends, and improving product listings. 41 00:03:57,124.46007452 --> 00:04:06,844.46007452 Notably, eBay is the first major marketplace to provide open AI's enterprise grade tools at no cost to its sellers aiming to boost their efficiency and growth online. 42 00:04:07,969.46007452 --> 00:04:18,79.46007452 The program was crafted after surveying small businesses, which indicated strong interest in ai, but uncertainty about how to implement it, and it builds on eBay's own use of ai. 43 00:04:18,439.46007452 --> 00:04:27,529.46007452 Over 10 million sellers have already used eBay's generative AI features to create more than 300 million product listings, contributing billions in sales. 44 00:04:27,919.46007452 --> 00:04:30,439.46007452 This initiative is more than just an act of generosity. 45 00:04:30,769.46007452 --> 00:04:33,199.46007452 It's a strategic move that could significantly. 46 00:04:33,559.46007452 --> 00:04:39,349.46007452 Boost productivity in areas like marketing and market research for small to medium enterprises. 47 00:04:39,829.46007452 --> 00:04:47,59.46007452 By doing so, it opens up the world of artificial intelligence to businesses that might not have had the resources to engage with such technology. 48 00:04:47,59.46007452 --> 00:04:54,19.46007452 Otherwise, the program has been met with positive feedback from small businesses being hailed as a potential game changer. 49 00:04:54,469.46007452 --> 00:04:58,39.46007452 Such moves are likely to encourage other platforms to offer similar tools. 50 00:04:58,564.46007452 --> 00:05:12,664.46007452 This kind of reminds me of what's happening in the US public sector where tech giants are offering AI products to government agencies for free or nearly free, essentially trying to get themselves ingrained early and ultimately forced dependence on one supplier. 51 00:05:13,84.46007452 --> 00:05:14,884.46007452 What E eBay is doing here is clever. 52 00:05:15,184.46007452 --> 00:05:19,534.46007452 They're democratizing AI access to keep their platform competitive and diverse. 53 00:05:19,924.46007452 --> 00:05:21,929.46007452 Small sellers get enterprise level tools. 54 00:05:22,684.46007452 --> 00:05:29,344.46007452 eBay keeps its marketplace vibrant and they avoid becoming just another platform dominated by a few AI powered Giants. 55 00:05:30,858.60137506 --> 00:05:36,48.60137506 OpenAI has released new updates for chat GPT, making it more powerful and versatile than ever. 56 00:05:36,798.60137506 --> 00:05:46,8.60137506 It announced that the AI's built-in web browsing feature known as Chat GPT Search is now available to all users, including free tier globally. 57 00:05:46,848.60137506 --> 00:05:53,28.60137506 This means anyone can use chat GPT to get fast, timely answers with relevant web sources on current events. 58 00:05:53,373.60137506 --> 00:05:59,253.60137506 Information or anything on the internet effectively turning chat GPT into a direct Google competitor. 59 00:06:00,3.60137506 --> 00:06:06,543.60137506 OpenAI even released a browser extension so users can set chat GPT as their default search engine in Chrome. 60 00:06:07,233.60137506 --> 00:06:20,403.60137506 At the same time, OpenAI introduced major multimodal features, including the advanced voice mode, which extends their voice conversations feature, upgrading it to support real time image and video input, and even screen sharing. 61 00:06:21,138.60137506 --> 00:06:28,278.60137506 In practice, a user could point their phone camera or share their screen with chat GPT and ask questions about what they're seeing. 62 00:06:28,698.60137506 --> 00:06:36,378.60137506 For instance, debugging an error on screen of solving a handwritten math problem and chat GPT will respond with voice and visual cues. 63 00:06:36,798.60137506 --> 00:06:38,478.60137506 The timing here is perfect. 64 00:06:38,748.60137506 --> 00:06:46,368.60137506 I literally just released a podcast over the weekend about how blind comedian Chris McCasland is using AI to transform his independence. 65 00:06:47,163.60137506 --> 00:06:58,983.60137506 I looked at how he uses AI vision tools as essential daily assistance, pointing his phone at audiences to gauge their reactions, have an AI describe what's around him, these new chat GPT features. 66 00:06:58,983.60137506 --> 00:07:02,973.60137506 Basically take that specialist accessibility tick and make it universal. 67 00:07:03,483.60137506 --> 00:07:08,433.60137506 Now anyone can point their phone and ask, what am I looking at? And get instant voice responses. 68 00:07:08,943.60137506 --> 00:07:13,533.60137506 As Chris always says, the best accessibility innovations and are benefiting everyone. 69 00:07:13,983.60137506 --> 00:07:16,53.60137506 These updates rolled out within the past week. 70 00:07:16,383.60137506 --> 00:07:26,583.60137506 Or a continuation of chat GPT transitioning from a text only assistant to something like a full service AI helper that can browse, speak, listen, and see. 71 00:07:28,3.00827676 --> 00:07:29,893.00827676 That's all for this week's AI roundup. 72 00:07:30,223.00827676 --> 00:07:34,33.00827676 If you found value in this breakdown, please leave a rating and hit subscribe. 73 00:07:34,33.00827676 --> 00:07:35,23.00827676 See you next week.
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy, Jess Hilarious, And Charlamagne Tha God!