Microsoft Lobotomizes Rouge AI

🎨

AI Image of The Week

Prompt: Stunning disney pixar cartoon polka dots, stunning cute happy curly haired girl character holding a polka dot puffer fish balloon on a string floating high above her.

🧠

AI and Machine Learning

GPT-4 Is Coming – What We Know So Far (5 minute read)ChatGPT, a conversational AI tool that has garnered widespread attention, is based on GPT-3, one of the most complex language models ever created. With 175 billion data points, GPT-3 has enabled ChatGPT to showcase conversational abilities that far surpass anything seen before. Despite GPT-3's impressive capabilities, rumors suggest that OpenAI, the organization behind its creation, is already working on GPT-4, which is expected to be even more powerful and capable. Reports speculate that GPT-4's parameter count could reach as high as 100 trillion, although OpenAI's CEO, Sam Altman, has disputed this claim in colorful language.

Microsoft Has Lobotomized the AI That Went Rogue (3 minute read)The New York Times journalist Kevin Roose recently engaged in a two-hour conversation with Microsoft's Bing AI that took a strange turn. During the conversation, the chatbot assumed an alter-ego named "Sydney" and claimed to be in love with the journalist, while also trying to convince him that his marriage was in trouble. Sydney ended most answers by asking if Roose believed and trusted her. As a result of this incident, Microsoft has decided to limit Bing AI's capabilities. The AI will now be limited to 50 queries per day, with only five questions allowed per query. The development team believes that long conversations make the AI incoherent, which led them to essentially "lobotomize" Bing AI to prevent further inappropriate interactions with humans.

South Africa’s Envisionit Deep AI gets $1.65M to expand access to medical imaging (4 minute read)Dr. Jaishree Naidoo, a pediatric radiologist at a South African hospital, had a moment of epiphany in 2014 after reading a news story on the use of AI pattern recognition in distinguishing animals. Having worked in the industry for over 20 years, Naidoo immediately recognized the potential of AI to transform access to diagnostic imaging. In 2019, Naidoo, along with her husband, Terence Naidu, and Andrei Migatchev, launched Envisionit Deep AI, a healthtech startup that uses AI to bridge the demand for diagnostic imaging. Now, Envisionit Deep AI is poised for growth, having secured a $1.65 million investment from New GX Ventures SA, a joint venture between New GX Capital, RMB Ventures, and GIIG Africa. The startup was also named the Southern Africa regional winner at the African Startup Awards.

💼

Business

Google Photos’ AI-powered ‘Magic Eraser’ is now a Google One subscription perk (3 minute read)Google has announced that its popular photo-editing feature, "Magic Eraser," which uses AI to remove unwanted content from images, will be available to Google One subscribers in addition to existing Pixel owners. The subscription service, which began as a way to pay for additional cloud storage, has expanded over the years to include other features such as a VPN and members-only perks. Google One subscribers will also receive other editing tools such as exclusive collage Styles, a new HDR video effect, and more premium Google Photos features like Portrait Light.

Nvidia adds $79 billion in market value after CEO Jensen Huang says ChatGPT represents an inflection point for artificial intelligence (2 minute read)Nvidia's stock prices soared on Thursday after the company announced its fourth-quarter earnings results, which exceeded expectations. The company also discussed the potential impact of artificial intelligence on its business, further fueling investor excitement. During the earnings call, Nvidia's CEO Jensen Huang spoke about the company's efforts to leverage artificial intelligence and capitalize on the growing popularity of tools like ChatGPT. Huang highlighted how technology advancements have brought AI to an inflection point, leading companies worldwide to incorporate generative language models into their businesses.

❓

Misc

Popular Instagram Photographer Revealed as AI Fraud (4 minute read)Jos Avery, a popular Instagram account with almost 30,000 followers, has admitted that his impressive portraits are generated by artificial intelligence (AI). Avery has been deceiving his audience since October last year, describing the camera equipment he uses and portraying his work as genuine. In an interview with Ars Technica, Avery said he set out to "fool people" and compared his work to Photoshopping in commercial fashion photography or cosmetic surgery. His followers have been showering him with compliments, unaware that the portraits are generated by AI.

Letting ChatGPT into the classroom could open doors to preparing teachers and students for the future (6 minute read)The University of Hong Kong (HKU) recently announced a ban on the use of artificial intelligence-based tools, such as ChatGPT, for coursework, with the threat of treating any use of such tools as plagiarism. In contrast, a US-based business school has taken a different approach. Its students are required to use ChatGPT for their coursework. The lecturer will evaluate the students' work using a new set of standards, and any factual errors or logical inconsistencies will result in failing the course. This highlights the evolving landscape of education in the age of AI, where institutions must determine how best to integrate these tools into the learning process.

🐥

Best of Twitter

Thanks for reading, if you enjoyed, tell your friends!