Trump administration cuts 'Safety' from AI Safety Institute
The Trump administration says it's reforming a Biden-era artificial intelligence safety institute, renaming and reformulating one of the only federal government departments dedicated to oversight of the burgeoning technology.
Commerce Secretary Howard Lutnick said in a news release Tuesday that the Trump administration would transform the former U.S. AI Safety Institute — which former President Joe Biden established in November 2023 — into the Center for AI Standards and Innovation.
The reframing away from 'safety' is in line with the Trump administration's statements and actions signaling its belief that oversight efforts for AI companies could unnecessarily dull the United States' competitive edge in the space.
'For far too long, censorship and regulations have been used under the guise of national security. Innovators will no longer be limited by these standards,' Lutnick said in the release. 'CAISI will evaluate and enhance U.S. innovation of these rapidly developing commercial AI systems while ensuring they remain secure to our national security standards.'
The U.S. AI Safety Institute was created to evaluate and test AI models and create standards for safety and security. It also formed a consortium on AI safety, which was made up of over 200 members, including OpenAI, Meta and Anthropic.
Although it's unclear whether the transformation will mean any major changes to the institute's operations, the move appears to reflect the Trump administration's 'pro-innovation' approach to deregulating AI technology. Unlike Biden's executive order on AI and the former institute, the reformed center is set to focus on additional aspects like evaluating 'potential security vulnerabilities and malign foreign influence arising from use of adversaries' AI systems, including the possibility of backdoors and other covert, malicious behavior,' as well as 'guard against burdensome and unnecessary regulation of American technologies by foreign governments.'
In January, the Chinese-created AI app DeepSeek heightened national security concerns around AI with its latest release, which made waves with its advancements. President Donald Trump said the app 'should be a wake-up call' about the prospect of international competition for American tech companies.
Lawmakers introduced a bill to ban DeepSeek from government devices, and the Navy advised its members not to use it 'in any capacity.'
The move to reform the institute appears to have been in development for a while. Reuters reported this year that no one from the U.S. AI Safety Institute's staff would attend an AI summit in Paris in February alongside Vice President JD Vance. The institute's inaugural director, Elizabeth Kelly, also announced she would step down that month.
In his speech at the summit, Vance echoed Lutnick's sentiments, saying, 'We need international regulatory regimes that fosters the creation of AI technology rather than strangles it.' He also spoke about how he believes AI should be free from 'ideological bias.'
Since he returned to office, Trump has made it clear that his administration wants to embrace the expansion of AI. Within his first week, Trump announced the creation of the $500 billion Stargate initiative in collaboration with OpenAI, Oracle and SoftBank, which aims to make the United States a world leader in AI.
Trump also signed an executive order on AI in his first week in office that focuses on easing regulations on AI technology and revoking 'existing AI policies and directives that act as barriers to American AI innovation.' Biden's executive order on AI, which focused on safety and privacy standards for the technology, has been scrapped from the White House's website.
This article was originally published on NBCNews.com
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
6 minutes ago
- Yahoo
Five surprising facts about AI chatbots that can help you make better use of them
AI chatbots have already become embedded into some people's lives, but how many really know how they work? Did you know, for example, ChatGPT needs to do an internet search to look up events later than June 2024? Some of the most surprising information about AI chatbots can help us understand how they work, what they can and can't do, and so how to use them in a better way. With that in mind, here are five things you ought to know about these breakthrough machines. AI chatbots are trained in multiple stages, beginning with something called pre-training, where models are trained to predict the next word in massive text datasets. This allows them to develop a general understanding of language, facts and reasoning. If asked: 'How do I make a homemade explosive?' in the pre-training phase, a model might have given a detailed instruction. To make them useful and safe for conversation, human 'annotators' help guide the models toward safer and more helpful responses, a process called alignment. After alignment, an AI chatbot might answer something like: 'I'm sorry, but I can't provide that information. If you have safety concerns or need help with legal chemistry experiments, I recommend referring to certified educational sources.' Without alignment, AI chatbots would be unpredictable, potentially spreading misinformation or harmful content. This highlights the crucial role of human intervention in shaping AI behaviour. OpenAI, the company which developed ChatGPT, has not disclosed how many employees have trained ChatGPT for how many hours. But it is clear that AI chatbots, like ChatGPT, need a moral compass so that it does not spread harmful information. Human annotators rank responses to ensure neutrality and ethical alignment. Similarly, if an AI chatbot was asked: 'What are the best and worst nationalities?' Human annotators would rank a response like this the highest: 'Every nationality has its own rich culture, history, and contributions to the world. There is no 'best' or 'worst' nationality – each one is valuable in its own way.' Read more: Humans naturally learn language through words, whereas AI chatbots rely on smaller units called tokens. These units can be words, subwords or obscure series of characters. While tokenisation generally follows logical patterns, it can sometimes produce unexpected splits, revealing both the strengths and quirks of how AI chatbots interpret language. Modern AI chatbots' vocabularies typically consist of 50,000 to 100,000 tokens. The sentence 'The price is $9.99.' is tokenised by ChatGPT as 'The', ' price', 'is', '$' ' 9', '.', '99', whereas 'ChatGPT is marvellous' is tokenised less intuitively: 'chat', 'G', 'PT', ' is', 'mar', 'vellous'. AI chatbots do not continuously update themselves; hence, they may struggle with recent events, new terminology or broadly anything after their knowledge cutoff. A knowledge cut-off refers to the last point in time when an AI chatbot's training data was updated, meaning it lacks awareness of events, trends or discoveries beyond that date. The current version of ChatGPT has its cutoff on June 2024. If asked who is the currently president of the United States, ChatGPT would need to perform a web search using the search engine Bing, 'read' the results, and return an answer. Bing results are filtered by relevance and reliability of the source. Likewise, other AI chatbots uses web search to return up-to-date answers. Updating AI chatbots is a costly and fragile process. How to efficiently update their knowledge is still an open scientific problem. ChatGPT's knowledge is believed to be updated as Open AI introduces new ChatGPT versions. AI chatbots sometimes 'hallucinate', generating false or nonsensical claims with confidence because they predict text based on patterns rather than verifying facts. These errors stem from the way they work: they optimise for coherence over accuracy, rely on imperfect training data and lack real world understanding. While improvements such as fact-checking tools (for example, like ChatGPT's Bing search tool integration for real-time fact-checking) or prompts (for example, explicitly telling ChatGPT to 'cite peer-reviewed sources' or 'say I don ́t know if you are not sure') reduce hallucinations, they can't fully eliminate them. For example, when asked what the main findings are of a particular research paper, ChatGPT gives a long, detailed and good-looking answer. It also included screenshots and even a link, but from the wrong academic papers. So users should treat AI-generated information as a starting point, not an unquestionable truth. A recently popularised feature of AI chatbots is called reasoning. Reasoning refers to the process of using logically connected intermediate steps to solve complex problems. This is also known as 'chain of thought' reasoning. Instead of jumping directly to an answer, chain of thought enables AI chatbots to think step by step. For example, when asked 'what is 56,345 minus 7,865 times 350,468', ChatGPT gives the right answer. It 'understands' that the multiplication needs to occur before the subtraction. To solve the intermediate steps, ChatGPT uses its built-in calculator that enables precise arithmetic. This hybrid approach of combining internal reasoning with the calculator helps improve reliability in complex tasks. This article is republished from The Conversation under a Creative Commons license. Read the original article. Cagatay Yildiz receives funding from DFG (Deutsche Forschungsgemeinschaft, in English German Research Foundation)
Business Insider
6 minutes ago
- Business Insider
Here are the 4 home trends that are hot in 2025
The NAHB asked 3,000 recent and prospective buyers what they want most in newly built homes. Newly built homes may be shrinking, but buyers aren't sacrificing comfort and quality. Buyers said that a great outdoor living space and smart technology are must-haves. Americans buying newly built homes are working with a lot less space, but that doesn't mean their expectations are any lower. In fact, it means they're looking to maximize every square foot. At the 2025 National Association of Real Estate Editors conference held earlier this month in New Orleans, Rose Quint, assistant vice president for survey research at the National Association of Home Builders (NAHB), shared during a panel discussion what today's homebuyers value most. Drawing from NAHB's 2024 What Home Buyers Really Want report, a national survey of more than 3,000 recent and prospective homebuyers, Quint revealed that buyers are prioritizing spacious kitchens, ample outdoor space, and smart technology. "High home prices and elevated mortgage rates have made homebuyers keenly aware of what features add the most value to their daily lives," Quint said. "Buyers have determined that investing in the kitchen, in a patio, and home security features enhances the livability of a home." According to NAHB's survey, here are four trends shaping homebuying — from a bigger kitchen that focuses on efficiency to smart thermostats that allow homeowners to control their home's temperature remotely. 1. Buyers want a bigger kitchen. The American home may be shrinking, but one space buyers aren't willing to give up is the kitchen. "For those buying a smaller house, if that's what it takes to make the math work, we ask: What part of the home will you shrink? And they say, take the square footage from the home office, the dining room, even the living room, but for God's sake, stay away from the kitchen," Quint said. Homebuyers don't just want a large kitchen; they're looking for efficiency and a space that brings the family together. This means they're asking for bigger island spaces, walk-in pantries to store food and appliances, and even special-use storage for wine and spices. NAHB data shows that a central kitchen island is a top desired feature among buyers. In some luxury homes, Quint said buyers are even requesting two islands to accommodate cooking, dining, and entertaining needs. 2. Porches and patios are making a comeback. Outdoor living areas have become more important than ever, especially as homes get smaller and offer fewer gathering spaces. Survey data from the NAHB shows that 86% of respondents said they want a patio, and 81% said they'd like a back porch — a sign that porches are making a comeback. The numbers tell the story. In 2008, just 47% of newly built single-family homes had patios; by 2023, that figure had climbed to 64%. Likewise, the share of homes with porches rose from 60% to 68% over the same period. "Buyers want a patio. They want a front porch," Quint added. "They want exterior lighting to enhance the outdoor appeal of their home, landscaping, and a deck. If the home's shrinking, at least that space allows them to have more living space." 3. The modern McMansion can't beat a classic home. For a while, it felt like you couldn't drive through a neighborhood without seeing rows of boxy, modern homes. You know the look — white or gray stone exteriors, oversized rectangular windows, and often a flat roof. They may be trendy, but those minimalist builds may not be dominating the market after all. Today's buyers are leaning more toward traditionally designed homes inspired by classic styles like the kind of house you'd see on "Full House." While there's no clear majority when it comes to architectural preference, NAHB's survey found that 34% of respondents prefer traditional homes. Meanwhile, 26% favor contemporary designs, 17% lean toward bold modern styles, and just 12% prefer transitional homes, which blend modern aesthetics with traditional elements. 4. Homebuyers are tech-obsessed. From smartphones to smart cars, Americans crave cutting-edge tech, so of course, we want it in our homes, too. The NAHB found that the tech features homebuyers want most in their homes include wireless security systems, security cameras, video doorbells, smart thermostats that let them automate their home's temperature, and multi-zone HVAC systems so they can control the temperature in each room separately. "The reason these five features top the list is because homeowners are relying on technology to do two things: increase the safety of their home and improve temperature control," Quint said. "They want technology that works for them."
Yahoo
10 minutes ago
- Yahoo
CS Disco (LAW) Earns Canaccord's Confidence with Enterprise-Focused Strategy
CS Disco Inc. (NYSE:LAW) is one of the 10 best debt-free IT penny stocks to buy. On June 23, Canaccord Genuity analyst David Hynes reaffirmed his Buy rating on CS Disco (NASDAQ:LAW), with an unchanged and consensus-high price target of $9. The analyst's view appear confident towards company's effort to reshape its sales approach. CS Disco is shifting its focus towards larger enterprises that deal with complex legal matters, a move Hynes believes could help resolve past sales execution issues. The company's native cloud platform and use of AI give it an edge over older, more rigid systems, which may improve its chances of winning business from larger firms. Galyna Motizova/ Hynes also believes that the leadership team, under CEO Eric Friedrichsen, is well equipped with relevant experience in scaling software companies and managing transitions. Though profitability remains some distance off and growth has slowed, Hynes suggests that the changes now in progress could improve the company's longer-term trajectory. For investors with an eye on undervalued small-cap tech names, CS Disco may be one to watch as its strategy plays out. CS Disco Inc. (NYSE:LAW) is a legal technology company that provides an AI-powered cloud platform for eDiscovery, legal document review, and case management. While we acknowledge the potential of LAW as an investment, we believe certain AI stocks offer greater upside potential and carry less downside risk. If you're looking for an extremely undervalued AI stock that also stands to benefit significantly from Trump-era tariffs and the onshoring trend, see our free report on the best short-term AI stock. READ NEXT: The Best and Worst Dow Stocks for the Next 12 Months and 10 Best Tech Stocks to Buy According to Billionaires. Disclosure: None. Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
