🗞 ChatGPT Gets News, OpenAI Eyes Chrome, and Character.AI Debuts AvatarFX

🔎 The Latest on the AI Frontier:

• Washington Post Teams with OpenAI to Bring Journalism to ChatGPT Results

• OpenAI Signals Strong Interest in Acquiring Chrome Browser Amid Antitrust Case

• Character.AI Releases AvatarFX for Single-Image Video Animation with Synchronized Speech

• Research Identifies Four Consumer Personality Types Interested in AI Assistants

• Musicians Deploy "Adversarial Noise" Technology to Shield Creative Work from AI Training

• Microsoft Unveils BitNet b1.58: CPU-Friendly AI Model that Doesn't Require GPUs

• Other news you might find interesting

📰 The Washington Post forms strategic partnership with OpenAI to feature its journalism in ChatGPT search responses

• ChatGPT will display summaries, quotes, and links to The Post's original reporting across politics, global affairs, business, and technology topics, with clear attribution and direct links to full articles.

• The partnership aims to make reliable, factual information more accessible to ChatGPT's 500 million weekly users, especially on complex or fast-moving topics where timely reporting matters most.

• This collaboration builds on The Post's AI initiatives including Ask The Post AI, Climate Answers, and Haystacker, reinforcing its commitment to expanding journalism discoverability through AI tools while remaining "LLM-agnostic."

🌐 OpenAI expresses strong interest in acquiring Google Chrome browser if antitrust regulators force a divestiture

• ChatGPT head of product Nick Turley testified in court that OpenAI would purchase Chrome "in a heartbeat," viewing the browser's 4 billion users (approximately 67% market share) as a massive distribution opportunity to deliver an "AI-first" browsing experience.

• The testimony revealed OpenAI previously approached Google about using its search technology within ChatGPT after experiencing "significant quality issues" with its current search provider, but Google declined the partnership request in August.

• Judge Amit Mehta is weighing remedies including potential Chrome divestiture by August after ruling Google maintains a search monopoly, while Google maintains Chrome cannot survive independently and plans to appeal the monopoly ruling.

🎬 Character.AI launches AvatarFX model that turns a single image into realistic video with synchronized speech, body, and face movements

• The new model uses a flow-based diffusion transformer pipeline paired with Character.AI's proprietary TTS engine to generate motion that tracks audio frame-by-frame, overcoming limitations of competitors like Pika and Runway Gen-3 by supporting longer clips and maintaining consistency.

• CAI+ subscribers will receive priority access with broader availability planned "over the coming months," with this launch following Google's reported $2.7 billion licensing-and-talent deal that integrated Character.AI's founders into the Gemini effort.

• Character.AI, founded in 2021 by ex-Google researchers Noam Shazeer and Daniel De Freitas, now attracts approximately 100 million monthly visits and raised $150 million in a 2023 Series A at a $1 billion valuation.

🤖 New research identifies four distinct personality types interested in consumer AI agents, revealing targeted opportunities for businesses

• Salesforce's survey of 2,552 US consumers found four key personas: the Smarty Pants (43%) who values informed decision-making, the Minimalist (22%) who prioritizes simplification, the Life-Hacker (16%) who seeks efficiency, and the Tastemaker (15%) who desires personalized recommendations.

• Consumer enthusiasm for AI agents spans multiple use cases with 44% interested in personal assistance (rising to 70% among Gen Z), 24% comfortable with AI shopping for them, 44% wanting help finding suitable jobs, and 43% seeking health and wellness management.

• The research suggests businesses should align agent development with these personality profiles to deliver tailored experiences, as 65% of consumers look to AI agents to improve decision-making and simplify their lives in an era where experience is as important as products and services.

🛡️ Musicians and artists develop "adversarial noise" technology to protect creative work from AI training without consent

• Activist Benn Jordan has combined "HarmonyCloak" (developed by University of Tennessee researchers) with his own "Poisonify" system to create audio that sounds normal to humans but "makes an acid casualty of artificial intelligence," breaking AI's ability to recognize harmony, rhythm, and instruments.

• This protective technology emerges amid growing tensions over AI training data usage, highlighted by Jack Dorsey and Elon Musk's recent controversial statements about wanting to "Delete all IP law" despite their own empires being built on intellectual property protections.

• The article argues that without proper intellectual property frameworks, tech companies remain vulnerable to these adversarial techniques, as this "digital LSD" can effectively poison AI models when embedded in training data, potentially undermining entire business models.

💡 Microsoft introduces BitNet b1.58, an AI model that runs efficiently on regular CPUs without needing GPUs

• Microsoft Research scientists have developed a groundbreaking "1-bit architecture" that stores and processes weights using only three values (-1, 0, and 1), eliminating the need for energy-intensive floating-point operations and allowing the model to run on standard CPUs using simple addition and subtraction.

• Testing shows the new model can match or even outperform similarly-sized GPU-based models while using significantly less memory and energy, potentially enabling users to run AI chatbots locally on personal computers or phones without requiring massive data centers.

• This innovation could dramatically reduce energy consumption for AI applications while also improving privacy by allowing AI processing to happen locally without an internet connection, supported by a custom runtime environment called bitnet.cpp designed to optimize the 1-bit architecture.

More news you might find interesting:  

• OpenAI CEO Sam Altman steps down as board chair of nuclear startup Oklo to enable potential strategic partnerships.

• Apple removes "available now" claim for Apple Intelligence features after advertising watchdog intervention.

• Israeli students to receive personal AI tutors through partnership between eSelf and major K-12 textbook publisher.

• DOJ challenges Google's AI strategy in remedies phase of major antitrust trial, seeking to prevent search monopoly extension.

• MIT Technology Review explores how artists are finding innovative ways to collaborate with AI.

• Elon Musk announces significant reduction in his involvement with government cost-cutting DOGE initiative starting May 2025.

• China's export restrictions on rare earth magnets impact Tesla's Optimus robot production amid escalating US-China trade tensions.

• EU Commission imposes hefty fines on Apple (€500 million) and Meta (€200 million) for Digital Markets Act violations.

• MIT researchers create "periodic table of machine learning" that connects classical algorithms and reveals new possibilities.

Have any feedback? Send us an email