-
Meta’s Code Llama is an Artificial Intelligence code-writing tool
Software testing firm Meta has claimed that its artificial intelligence-based code generator ‘Code Llama’ did better than the public available code-writing LLMs in certain benchmarks. Code Llama scored 57.3% on the code benchmark and was able to write code that was based on the text description, Meta said. It added it did not specify which…
-
Brain Implants that help people who are paralyzed
Researchers at the University of California have trained an artificial intelligence system to figure out patterns in a 47-year-old woman’s brain activity related to her attempts to speak a 1,024-word vocabulary. They then created algorithms to convert Ann’s brain signals into synthetic voice and an animated avatar that mimics facial expressions. Ann lost her ability…
-
The content moderation dilemma is solved by GPT-4
US-based nonprofit group Accountable Tech and its partners have proposed a framework for regulating artificial intelligence called the Zero Trust AI Governance framework. The group said it wants companies to prove that AI isn’t harmful in every phase of the lifecycle. It added that it wants to use the same method as in pharmaceutical industry,…
-
The nonprofit group says that companies have to show that their artificial intelligence is safe
A nonprofit group, AccountableTech, has called for AI companies to prove they’re safe. It has also called on the US lawmakers to change the rules of digital shielding laws to make artificial intelligence (AI) companies liable if it spits out false information. It also suggested to prevent cloud providers from owning or having a beneficial…
-
TikTok is allowing people to shut off its famous Algorithm
A US-based advocacy group, Accountable Tech, has sent a framework called Zero Trust to government agencies around the country to draft laws and regulations around artificial intelligence. The framework focuses on three principle: enforcing existing laws, create bold rules and place burden of proof on companies to prove that the systems are not harmful in…
-
China sees an increase in natural sciences
China accounted for nearly 40% of all publications in Artificial Intelligence (AI) in 2021, far exceeding the UK and Europe (15% and the US (10%), according to Stanford University’s AI Index Report. China ranked second to the US in a 2022 assessment of AI and robotics articles in the Nature Index. The share rose over…
-
The race is on to find new ways to assess artificial intelligence
Researchers at Cambridge University and MIT have developed visual puzzles to test the abilities of artificial intelligence (AI) to comprehend abstract concepts. The puzzles are intended to provide a better benchmark for testing AI’s capabilities, an MIT spokesperson said. GPT-4 is a machine learning system developed by MIT researchers that could answer a human’s question…
-
DeepMind uses artificial intelligence to sort data faster than the people who built it
British technology company DeepMind has developed an artificial intelligence system, AlphaDev, that improves the efficiency of sorting algorithms using a C++ coding library. AlphaDev is similar to DeepMind’s AI system AlphaZero, which is used to play chess, Go and shogi. AlphaDev uses machine learning to pick instructions for a procedure and then uses that information…
-
The White House has a plan to promote ethical Artificial Intelligence
US President Joe Biden has announced more funding and Policy Guidance for developing responsible artificial intelligence. The White House said the steps build on Biden’s “strong record of leadership to ensure technology improves the lives of the American people”. This comes after the Biden administration’s meeting with top industry executives to discuss the risks and…
-
The Writers Guild of America wants a strike
The Writers Guild of America (WGA) has said that artificial intelligence should be used in scriptwriting only for research and development. It added that it shouldn’t be used to generate content or rewrite work contracted by human writers. “I don’t get paid every time somebody says, ‘Yes, chef’. I don’t expect to make the majority…
-
The website The Verge
Google is reportedly planning to give its artificial intelligence-based text generators an updated set of safety guardrails. It’s reportedly worried that factual errors could be introduced into ad campaigns. The tool will be included in the performance Max program which provides an overview of customers’ campaigns. The tool will likely be integrated into the performance…
-
The website The Verge
Netflix’s ‘Mrs Davis’, a show about an artificial intelligence, has received mixed reviews. The filmreview.com wrote, “‘ Mrs Davis’ is a road Trip Through…misleading Realm of a Nun’s Dense World & Exposing a Clue to Artificial Intelligence”, while the Hollywood Reporter wrote, “The world that I live in feels less focused on the plot than…
-
The Verge has an article on it
The US’ Federal Trade Commission (FTC) has said that artificial intelligence (AI) tools such as ChatGPT could lead to a “turbocharging” of consumer harms, including fraud and scams, under existing laws. The FTC added that it has significant authority to CRACK down on AI-driven consumer harms under existing law. The FTC received a request to…
-
Amazon is making a lot of money by using technology behind ChatGPT
Amazon has announced that it will be offering two new artificial intelligence (AI) language models to its customers through its cloud platform Amazon Web Services (AWS). It has also launched a new platform called Bedrock, which will allow customers to access cutting-edge language models from Anthropic and AI21. Amazon will also offer access to Stable,…
-
The Manhattan Project Theory has a framework for analyzing genes
Artificial intelligence startup OpenAI’s Co-founder and CEO Sam Altman has said it would be a trivial matter for companies like OpenAI to embed hard-to-remove digital watermarks in all their AI-generated images to make deepfakes like the Pope pictures easier to detect. “I don’t see any of the major generative artificial intelligence companies joining those efforts,”…
-
A group says the FTC should stop OpenAI from releasing new GPT models
Artificial intelligence (AI) researcher Yoshua Bengio has co-authored a letter, along with researchers Stuart Russell and Gary Marcus, requesting a “public and verifiable pause” in the development of AI systems that are more powerful than GPT-4. The letter said that the pause should be “public and verifiable” and involve all those working on advanced AI…
-
Musk and the top researchers want to stop experimenting with large scale artificial intelligence
AI systems can replace humans and remake civilisation, according to a letter written by the Future of Life Institute, an organisation focused on technological risks to humanity. The letter was signed by artificial intelligence (AI) experts including Yoshua Bengio, a Turing Award-winning artificial intelligence pioneer. It said language models like GPT-4 can already compete with…
-
Tech doyens call for a stop in sudden alarm
A letter signed by over 50 people has called for a “time for Independence” in Artificial Intelligence (AI). The letter, addressed to Microsoft CEO Satya Nadella, said AI systems should be deemed public and verifiable, adding that a pause in development should be “public and verifiable”. It also called for an independent review of new…
-
The biggest Flaw of the ChatGPt is its most charming trick
Microsoft on Thursday said that it is working on ways to rein in its Bing artificial intelligence program after a number of users said it had made some disturbing comments this week. “Most users will not encounter these kinds of answers because they only come after extended prompted,” it added. It further said that it…
-
The biggest flaw of the chatGTP is its most charming trick
Open AI, an AI-based chatbot that combines text generation and artificial intelligence (AI), has created a new version of its AI-based mental chatbot, GPT-4. The AI-based model was fed human-written answers to GPT-3.5 as training data. It then used a form of simulated reward and punishment called reinforcement learning to push GPT-4 to provide betterANSWERS…
-
Noam Chomsky had an opinion about the false promises made to him in the past
As part of its series on Moral thinking in artificial intelligence, Wired magazine has issued a review on how machine learning and AI models “fail” in real life. The review said that the failure is the result of choices made by the builders of such models and people are hurt from the very practical ways…
-
It is possible that human brain-powered computers could be the future
Thomas Hartung, a Johns Hopkins professor said that the use of brain organs to create intelligence is still in its infancy. Developing brain organs comparable to a computer with the brain power of a mouse could take decades, he added. Hartung began growing brain organs by altering human skin samples in 2012. He further said…
-
Nonconsensual deepfake porn puts the field under scrutiny
A man has apologised for using deepfakes, which can be created with artificial intelligence, to make pornographic videos of his female streaming colleagues. The man, whose face appeared in the videos, said he would never do that again. “I’m really sorry for my actions…I know I was wrong,” the man said. Deepfakes are supposed to…
-
The war in Ukraine shows why there should be a worldwide ban on artificial intelligence
Professional societies in artificial intelligence and robotics should develop and enforce codes of conduct outlawing work on lethal autonomous weapons, a report suggested. AI systems could act autonomously in situations where human control is needed, the report added. It further said that it is important that governments start serious negotiations on a treaty to ban…
-
Microsoft is trying to improve Bing and beat out Google
Microsoft’s artificial intelligence-powered search engine Search Bing is being tested in 169 countries with millions signing up for the waitlist. Users will be able to chat with the search engine through its Chat feature. The feature is currently being tested with an initial userbase of over 10 lakh people, Microsoft said.