- What's next? TwitterGPT? A recently reported purchase has some wondering if Elon Musk is planning to build a ChatGPT rival.
- AI bounty hunters wanted. OpenAI launches “Bug Bounty,” a program offering rewards of up to $20,000 for info on AI vulnerabilities.
- Meta AI launch. Meta AI introduces us to SAM (Segment Anything Model).
Despite recently calling for a six-month pause on giant AI experiments, Insider has reported that Elon Musk just procured around 10,000 graphics processing units for the platform that are “typically” used when working on giant AI models.
One of the sources Insider referenced revealed that Musk's artificial intelligence initiative at incorporates a sizable language model (LLM) — a form of AI trained on vast quantities of data, allowing it to generate content and text independently like ChatGPT. Twitter possesses extensive data sets that could be used to train an LLM. OpenAI previously accessed Twitter for ChatGPT training purposes until Musk found out and put a stop to it.
News that Twitter hired two engineers from DeepMind, Alphabet’s AI-research organization, was bandied about in early March. Reportedly, Igor Babuschkin will serve as a senior director of engineering at Twitter and Manuel Kroiss will serve as the new senior director of software engineering.
So, it appears that Twitter is continuing to invest in AI research and development despite criticism from Elon Musk who once said, “One of the biggest risks to the future of civilization is AI,” and who now may be building his own version of ChatGPT.
In other AI news...
Altman to Musk: Stop Tweeting and Start Loving
Speaking of ChatGPT, OpenAI Founder Sam Altman was recently interviewed on the Greymatter Podcast where he shared a few thoughts about his old pal, Elon.
It’s not all drama. Altman once viewed Musk as a hero and believes the two men both agree on the downside of Artificial General Intelligence (AGI) and the importance of getting safety right — to ensure a world “where people are much better off because AGI exists.”
“Elon is obviously attacking us some on Twitter right now on a few different vectors,” Altman said during the podcast. “And I have empathy because I believe he is, understandably, so really stressed about AGI safety. I'm sure there are some other motivations going on too, but that's definitely one of them.”
But Altman does wish Musk would be a tad bit nicer about it.
“I definitely grew up with Elon as a hero of mine. You know, despite him being a jerk on Twitter or whatever, I'm happy he exists in the world,” Altman said. “But I wish he would do more to look at the hard work we're doing to get this stuff right. A little bit more love.”
Related Article: OpenAI Addresses AI Safety, Google CEO Talks Bard Bust, More News
OpenAI’s Bug Bounty Program Offers up to $20K for Finding AI Vulnerabilities
OpenAI has announced the launch of “Bug Bounty,” a program offering rewards of up to $20,000 for researchers who find vulnerabilities in its AI systems. Created through a partnership with HackerOne, the program includes a list of AI projects that are eligible, along with details about the types of vulnerabilities that qualify for a reward. Researchers who find a low-severity vulnerability can earn $200, and the more severe the vulnerability, the higher the payout — those who identify “exceptional discoveries” may earn up to $20,000.
“The OpenAI Bug Bounty Program is a way for us to recognize and reward the valuable insights of security researchers who contribute to keeping our technology and company secure,” company officials said in a blog on the OpenAI website. “By sharing your findings, you will play a crucial role in making our technology safer for everyone.”
For more information about the program, including guidelines and rules for participation are accessible through the Bug Bounty Program page.
Meta AI Releases Segment Anything for Image and Video
Meta AI has launched its Segment Anything Model (SAM), a new AI tech that promises to “cut out any object, in any image, with a single click.”
According to company officials, SAM uses AI to automatically segment and label objects in images and videos, removing the need for manual labeling. Segment Anything combines computer vision and natural language processing (NLP) to understand the context of an image or video and identify objects within it.
The tool can identify multiple objects in a single image and can also differentiate between objects with similar visual characteristics. You can try the demo now.
AI Video of the Week: The Question Warren Buffet Wants to Ask AI
Warren Buffet admits to not understanding AI “at all,” but when the Berkshire Hathaway CEO joined CNBC's Squawk Box, he said he would like to ask it one question — how is it going to ruin the human race?”
Questions About Generative AI? Ask a Google Techspert
Douglas Eck, a senior research director at Google, offered up his expertise in Google’s latest blog, What is Generative AI. As someone with a background in literature and music, who is also at the forefront of AI, Eck said, “That combination of the technical and the creative puts (me) in a special position to explain how generative AI works and what it could mean for the future of technology and creativity.”
Eck addresses a wide variety of questions including what exactly is AI, how do language models really work and what does generative AI mean for creative fields and creativity in general.
“Today, generative AI could help you write a form letter. Tomorrow, it may overhaul your creative workflows and processes to free you up to solve completely new challenges with a new frame of mind,” Eck said. “Through collaboration and experimentation over time, we’ll uncover even more benefits from generative AI.”
AI News Presenter 'Fedha' Debuts in Kuwait
An AI-generated news presenter has made her debut in Kuwait. The virtual news anchor, named “Fedha,” is able to deliver the news in Arabic and English using natural language processing technology. According to an article by Al Jazeera, Abdullah Boftain, deputy editor-in-chief of the Kuwait Times, said the move “is a test of AI’s potential to offer new and innovative content. And, looking to the future, they may even give her a Kuwaiti accent and make her a regular contributor to the site’s Twitter account.
“Fedha is a popular, old Kuwaiti name that refers to silver, the metal,” Boftain told Al Jazeera. “We always imagine robots to be silver and metallic in colour, so we combined the two.”
Request for AI Policy Input
The National Telecommunications and Information Administration (NTIA) wants to know what policies should shape the AI accountability ecosystem, and they need your help. So, NTIA has issued a request for comments on artificial intelligence policies in the development of AI audits, assessments and certifications for AI systems.
"We’re going to need input from public interest groups, from industry, from our federal partners. That is all essential for this strategy to succeed," Alan Davidson, assistant secretary of communications and information for NTIA, said.
The input received will inform the NTIA's policy on AI and guide its efforts in promoting the development of AI in a manner beneficial to society. The deadline for submitting comments is by June 10, 2023, Comments will be made publicly available.
AI Tweet of the Week: Using AI to Help People With Disabilities
It may not look exactly like a dog, but this AI tech acts like a guide dog for the blind or visually impaired — using Google maps to get around. Named “TEFI,” it can tell the difference between people and objects, and it communicates with its user. Researchers at the Institute of Physical and Information Technologies (ITEFI) are planning to add sensors to enhance its abilities and help it assist in medical emergencies.
This #Robot acts like a guide dog— Ronald van Loon (@Ronald_vanLoon) April 10, 2023
by @wef#AI #ArtificialIntelligence #MI #Robotics #Engineering #Innovation #FutureOfWork #Tech
cc: @wil_bielert @patrickgunz_ch @johnlegere pic.twitter.com/dmX0fB1r9e