ScarJo slams OpenAI for voice clone

On Monday, mega-film star Scarlett Johansson forced OpenAI to shut down its flirty chatbot voice, “Sky”, for being “eerily similar” to her own. This was just one week after OpenAI’s triumphant demo of its new, more powerful model, GPT-4o. The demo quickly went viral, and “Sky’s” voice was on prominent display during several of the most eye-popping moments. Turns out, OpenAI had asked Johannson for permission to use her voice, and she’d said no. (Consent, guys! No means no! Why is this so hard for you?) The company made several lame excuses - they had used a paid voice actress, who had used her natural voice, etc. etc. Too bad CEO Sam Altman had teased the demo on Twitter/X with the word “Her”, the name of a sci-fi movie in which a man falls in love with an AI - who was voiced by Johannson. (The gun! She is smoking!) Faced with a PR catastrophe and the real possibility of a successful legal damages suit, OpenAI caved, and “paused” use of The Voice Named “Sky.”

ScarJo slams OpenAI for voice clone

Microsoft announces AI laptops

On Monday, Microsoft CEO Satya Nadella announced a major change in the design of all of its future laptops - they will include AI on the device, with a powerful dedicated neural processing unit (NPU) in addition to the traditional central processing unit (CPU.) This is part of Microsoft’s “AI Everywhere” strategic focus, which seeks to weave AI into every aspect of consumer apps and hardware. The onboard NPU will run small local AI models, which will speed up response time for many tasks, but will connect to large AI models in the cloud as needed. Microsoft is enforcing this new architecture on all of its major laptop hardware partners (Dell, HP, Lenovo, etc.) One new capability of this AI-centric design is “Recall”, a searchable “photographic memory” of everything you’ve done or viewed on the laptop. (Privacy alert! Danger, Will Robinson!) How will this work out? As Yogi Berra said, “It’s hard to predict - especially the future.”

Microsoft announces Copilot+ AI laptops

Google to add ads to AI-powered search

Google has been forced by its AI rivals to start delivering AI-generated answers to queries, rather than just a list of links with juicy ad revenue attached to each one. Many doomers said that this spelled the end of Google’s torrent of ad revenue, threatening the viability of the company. Google’s simple solution: it’s planning to put ads on the AI answers. Check and mate, doomers. The only thing doomed was the hopes of long-suffering internet users trying to get some small bit of information without being assaulted by advertising.

Google to add ads to AI-powered search

Microsoft offers to relocate 700+ employees from China

As US-China competition in AI heats up, Microsoft has tried to get out of the crossfire by offering to relocate 700-800 of its Chinese employees who work on AI or cloud services. Employees reportedly were given their choice of relocation to the US, Ireland, Australia, or New Zealand. China is home to Microsoft’s largest research and development center outside the US, employing several thousand Chinese workers.

Microsoft offers to relocate 700+ employees from China

AI trained on “Friends” sitcom can detect sarcasm

Sometimes it’s hard to know when someone’s joking. Scientists have been laboring for years to teach machines how to do this for us, with little progress until lately. But today’s powerful multimodal AI models, able to parse text, image, and tone of voice simultaneously, seem to be up to the task. A recent paper details how a multimodal AI was trained on a database of film clips from “Friends”, “The Big Bang Theory”, and other suchlike sitcoms, and was able to succeed in detecting sarcasm 75% of the time. Although this research would appear to be more worthy of the tongue-in-cheek IgNobel Prize than the actual Nobel, the scientists assure us that these discoveries have practical applications in teaching our ubiquitous chatbots to better understand what we really mean.

AI trained on “Friends” sitcom can detect sarcasm

Waymo says its robotaxis are making 50,000 paid trips a week

Self-driving rideshare company Waymo, a subsidiary of Google parent Alphabet, announced on Twitter/X that it was now making 50,000 paid trips per week across its 3 initial cities of Phoenix, Los Angeles, and San Francisco. Waymo’s slow and deliberate rollout of its robotaxi service seems to be doing better than rival Cruise, a subsidiary of General Motors, which has been plagued by accidents and mishaps, and even suspended operations for a while to fix safety issues.

Waymo says its robotaxis are making 50,000 paid trips a week

Unitree unveils $16,000 humanoid robot for the home

The competition to make the first commercially successful humanoid robot is fierce, involving legendary companies such as Tesla and Boston Dynamics, and scrappy startups like Figure. Now Unitree Robotics of Hangzhou China, known heretofore as a maker of quadruped industrial robots, is throwing its hat in the ring with its new 4-foot tall H1, which is smaller, cuter, and way more flexible than rivals, able to spin its shoulder and hip joints in all directions. And way less spendy - only $16,000 each. (Such a bargain!)

Unitree unveils $16,000 humanoid robot for the home.

ChatGPT spikes to 2 billion visits in May after GPT-4o demo

ChatGPT, already one of the most popular websites in the world, has seen its traffic soar ever since its viral demo of its latest model, GPT-4o. Daily traffic has nearly doubled, from around 50 million daily visits, to over 100 million daily visits on at least 2 days this month. At this pace, ChatGPT is on track to log over 2 billion visits in the month of May.

ChatGPT spikes to 2 billion visits in May after boffo GPT-4o demo

Cancer surgery took away her voice. AI gave it back.

Thanks to AI, 21-year-old Lexi Bogan has her voice back. Last August she had a life-threatening tumor removed from the base of her brain. Weeks on a ventilator after surgery damaged her vocal cords, and the voice of the former soprano in her high school chorus was reduced to a raspy whisper. Last month, she got an artificial replacement - a voice prosthesis if you will - in the form of an AI voice clone trained on 15 seconds of video from a cooking demonstration that she recorded in high school. Now her old voice is an app on her phone. Way better use of the technology than cloning ScarJo’s voice against her wishes.

Cancer surgery took away her voice. AI gave it back.

Israeli AI flags clashing medications

Polypharmacy can cause hospitalizations due to conflicts between a patient’s medications. A major Israeli health system is attacking this problem with an AI system that warns doctors of potential adverse interactions between medications their patient is taking. The system continually monitors patients’ lab results as well their various medications, and alerts physicians of patients at risk for hospitalization.

Israeli AI seeks to reduce hospitalizations by flagging clashing meds

Korean Ministry of Health to develop clinical AI system for ERs

The Korean Ministry of Health and Welfare has announced a major technology initiative for the nation’s ERs, which will include an AI-powered Clinical Decision Support System for predicting catastrophic outcomes such as cardiac arrest or sepsis.

Korean Ministry of Health to develop clinical AI system for ERs