A federal judge in Manhattan recently issued a partial ruling in favor of Palantir Technologies against former employees who founded the AI firm Percepta. The judge blocked former Palantir engineer Joanna Cohen from violating a confidentiality agreement. However, the court did not prevent Percepta co-founders Hirsh Jain and Radha Jain from continuing to work at their new company or soliciting customers, despite finding they likely violated confidentiality and non-solicitation agreements. Palantir had alleged Percepta uses similar AI technology to make businesses more efficient, while the defendants maintain Percepta is a consulting firm.
In other AI developments, Google is launching a free AI Professional Certificate aimed at helping small businesses and individuals acquire valuable AI skills, including communication and content generation. This initiative includes free access to Google Workspace Business Standard for small businesses, building on the success of Google's foundational AI course, which is its most popular on Coursera. Meanwhile, Reddit is testing an AI-powered shopping search feature for a small group of U.S. users, displaying product carousels with prices and buying links directly from search results, drawing recommendations from user discussions.
The global focus on AI safety continues to grow. Britain's AI Safety Institute is leading international efforts, positioning the country as a key player in AI safety inspection, acknowledging both the benefits, like DeepMind's AlphaGenome for DNA understanding, and risks highlighted by Anthropic. Researchers at the University of Florida are proactively enhancing AI security by intentionally "breaking" systems from companies like Meta and Microsoft using a method called Head-Masked Nullspace Steering (HMNS) to identify and fortify against potential weaknesses.
Beyond safety, AI is finding diverse applications and sparking debate. Scientists at the University of New Hampshire utilized AI to discover 25 new high-temperature magnetic materials, a breakthrough that could reduce reliance on rare earth elements for electric vehicles. On the other hand, concerns are rising about AI's impact on online authenticity, with an author observing a decline in creativity as AI-generated content floods platforms, leading some, particularly Gen Z, to disengage. Yet, the Compliance Scorecard version 10 now uses AI for explainable compliance decisions, treating AI as decision support within controlled systems to meet regulatory demands.
Finally, the potential for international cooperation on AI risks is being explored, with the US and China recognizing shared threats that could foster collaboration on safety standards and responsible deployment, despite their geopolitical rivalry. In the business sector, luxury blanket retailer Minky Couture is adopting the Domo AI and Data Products Platform to integrate data from various departments, aiming for AI-powered insights to optimize inventory, enhance forecasting, and improve overall operational visibility.
Key Takeaways
- A federal judge partially sided with Palantir Technologies, ruling that former employees, including Hirsh Jain, likely violated confidentiality and non-solicitation agreements when founding their AI firm, Percepta.
- Google is offering a free AI Professional Certificate and Google Workspace Business Standard access to small businesses to bridge the AI skills gap, building on its most popular foundational AI course on Coursera.
- Reddit is testing an AI-powered shopping search feature for U.S. users, displaying product carousels with prices and buying links based on Reddit discussions.
- Britain's AI Safety Institute is leading global AI safety research, with DeepMind's AlphaGenome showing AI's benefits and Anthropic's warnings highlighting risks.
- University of Florida researchers are improving AI safety by intentionally "breaking" systems from companies like Meta and Microsoft using Head-Masked Nullspace Steering (HMNS) to identify weaknesses.
- AI has discovered 25 new high-temperature magnetic materials, potentially reducing the need for rare earth elements in electric vehicles and clean energy.
- Concerns are growing that widespread AI-generated content is diminishing internet creativity and authenticity, causing some users, particularly Gen Z, to disengage.
- The US and China are identified as having potential for cooperation on AI risks, despite geopolitical rivalry, to establish safety standards and responsible deployment.
- Compliance Scorecard version 10 now uses AI for explainable compliance decisions, treating AI as decision support within controlled systems to meet regulatory and client demands.
- Minky Couture is implementing the Domo AI and Data Products Platform to integrate data and gain AI-powered insights for optimizing inventory, forecasting, and operational visibility.
Judge stops ex-Palantir workers from hiring former colleagues
A federal judge in Manhattan has partially sided with Palantir Technologies in a lawsuit against former employees. The judge blocked former Palantir engineer Joanna Cohen from violating a confidentiality agreement. However, the judge did not stop the former employees from working at their new AI firm, Percepta, or soliciting customers. Palantir claims Percepta, founded by ex-employees Hirsh Jain and Radha Jain, uses similar AI technology to make businesses more efficient. The defendants argue Percepta is a consulting firm, not a software seller like Palantir.
Palantir wins partial victory against AI startup founded by ex-employees
A federal judge ruled that former Palantir employees likely broke confidentiality and non-solicitation agreements when starting their AI company, Percepta. While the judge did not stop them from working at Percepta, he agreed that they likely violated agreements. Palantir sued Percepta co-founder Hirsh Jain and two other former executives, claiming they recruited Palantir employees and misused confidential information. The defendants claimed victory, stating the court rejected Palantir's main claims and they can now focus on building Percepta.
AI finds new high-temperature magnets for electric vehicles
Scientists at the University of New Hampshire used AI to discover 25 new magnetic materials that remain magnetic at high temperatures. This breakthrough could lead to cheaper electric vehicles and clean energy technologies by reducing the need for rare earth elements. The AI analyzed a database of over 67,000 magnetic compounds to identify these promising materials. This discovery speeds up the search for sustainable magnets, which are crucial for many modern devices but often rely on costly and hard-to-obtain rare earth elements.
Reddit tests AI shopping search for users
Reddit is testing a new AI-powered search feature that helps users shop directly from search results. A small group of U.S. users will see product carousels with prices and buying links when searching for items. This feature pulls product recommendations from Reddit discussions, allowing users to tap on items for more details and purchase links. This move is part of Reddit's effort to integrate e-commerce capabilities into its platform. The company also noted a significant increase in users for its AI-powered Reddit Answers feature.
Google offers free AI certificate for small businesses
Google is launching the Google AI Professional Certificate to help people gain valuable AI skills for the job market. The certificate focuses on practical skills like communication, research, and content generation using AI tools. Google is offering free access to this certificate and Google Workspace Business Standard to all small businesses. This initiative aims to bridge the AI skills gap and ensure everyone can benefit from AI's economic growth. The company's foundational AI course is already its most popular on Coursera.
Britain leads the world in AI safety research
Britain is at the forefront of AI safety efforts with its AI Safety Institute, making it the closest the world has to an AI safety inspector. This comes as AI technology rapidly advances, with potential for both great benefit, like DeepMind's AlphaGenome for DNA understanding, and significant risks, as highlighted by Anthropic's warnings. The article suggests that while Britain has a leading role, continuous effort is needed to manage AI's dual nature effectively. The rapid development of AI necessitates ongoing attention to its safety and potential dangers.
Researchers break AI to make it safer
University of Florida researchers are intentionally breaking AI systems to improve their safety and security. By examining the internal workings of AI models, they identify weaknesses that could be exploited. This method, called Head-Masked Nullspace Steering (HMNS), helps developers build stronger defenses against misuse. The research focuses on stress-testing AI tools from companies like Meta and Microsoft to understand their limits. This proactive approach is crucial as AI becomes more integrated into critical sectors like healthcare and finance.
AI is harming the internet's creativity and authenticity
The author argues that the increasing use of AI-generated content is leading to a decline in creativity and authenticity online. They observe AI images and text flooding social media, making the internet feel less genuine. While acknowledging AI's value in areas like computer science and resume building, the author believes its widespread integration into everyday tools like browsers and emails is overwhelming. This trend, coupled with the pursuit of viral content, is causing many, especially Gen Z, to disengage from online platforms.
US and China can cooperate on AI risks
The US and China, despite their rivalry, can find common ground in managing the risks of artificial intelligence. History shows that shared threats can foster cooperation, as seen with the US and China aligning against the Soviet Union in 1972. AI presents a similar opportunity, as both nations recognize its potential for both progress and catastrophic risk. Cooperation is essential for establishing safety standards and responsible deployment, especially considering the significant energy demands of advanced AI infrastructure. Focusing solely on geopolitical competition overlooks the critical need for collaboration in this field.
New Compliance Scorecard uses AI for explainable decisions
Compliance Scorecard version 10 now uses AI to provide explainable compliance decisions within a controlled system. The platform ensures AI is only used with existing, validated context and controls, treating it as decision support rather than a conversational tool. This approach addresses demands from regulators and clients for AI-assisted compliance workflows that are explainable, auditable, and accountable. The system applies AI using real operational data, including tools, configurations, and policies, making its outputs transparent and trustworthy.
Minky Couture uses Domo AI for better business insights
Luxury blanket retailer Minky Couture is implementing the Domo AI and Data Products Platform to improve its operational visibility. The company aims to integrate data from various departments like inventory, supply chain, and marketing to become a more intelligent enterprise. Using Domo, Minky Couture will create data products to track performance, optimize inventory, and enhance forecasting. This partnership will enable AI-powered insights for better decision-making and support the company's philanthropic efforts.
Sources
- US judge blocks ex-Palantir staffers from poaching workers for new AI firm
- Palantir Gets Partial Win in Fight With Ex-Workers at AI Startup
- AI breakthrough could replace rare earth magnets in electric vehicles
- Reddit is testing a new AI search feature for shopping
- Practical AI skills for everyone
- Britain is the closest the world has to an AI safety inspector
- Breaking AI on purpose: How researchers are helping make artificial intelligence safer
- DUELING COLUMN: AI will be the death of the internet age
- Opinion | US and China can again find common ground
- Compliance Scorecard v10 delivers context-driven AI for explainable compliance decisions
- Minky Couture selects Domo AI, data products platform to boost operational visibility
Comments
Please log in to post a comment.