Weekly AI News: Grok Build, OpenAI Election Safeguards, CNN vs Perplexity, Samsung AI Chips, and More

This week in AI was not about one single big launch. The bigger story was how artificial intelligence is spreading across software development, media, smart homes, enterprise work, data centers, chips, and regulation. From xAI launching a coding agent to CNN using Perplexity, and from Samsung pushing faster AI memory to Wipro expanding agentic AI workflows, the week showed how AI is becoming both more useful and more contested.

Build Smarter Business Apps with AI


Prismetric develops custom AI solutions for software automation, coding agents, workflow intelligence, and enterprise product development.

Talk to AI Experts 

Monday, May 25: xAI Pushes Into Coding Agents

xAI launched Grok Build, an early beta coding agent and command-line tool for software engineering work. The tool runs from the terminal and is available to SuperGrok and X Premium Plus subscribers.

This matters because coding agents are becoming one of the most competitive areas in AI. These tools do more than suggest code. They can help plan work, edit files, review changes, and support more complex development tasks. For developers, this means AI is moving closer to becoming a practical coding partner rather than just an autocomplete tool.

Source: x.ai

Monday, May 25: OpenAI Signs Its First Major Media Partnership in Brazil

OpenAI announced a strategic content partnership with Grupo Folha and Grupo UOL, bringing journalism from Folha de S.Paulo and UOL into ChatGPT with attribution. The Brazilian media groups will also get access to OpenAI tools such as Codex, ChatGPT Enterprise, and the API.

The bigger point here is that AI companies and publishers are still trying to find a workable model for news. Some publishers are suing AI companies. Others are choosing licensing and partnership deals. For readers, the key question is whether AI answers can become more reliable while still supporting the original newsrooms that produce the reporting.

Source: openai

Wednesday, May 27: OpenAI Details Election Safeguards for 2026

OpenAI published a new update on election information and safeguards for 2026. The company said its focus is helping people access reliable information, supporting cyber defenders, and increasing AI transparency around elections.

This is important because 2026 is another major global election year, and generative AI is now widely available. AI can help voters understand issues, but it can also be used to create misleading content, fake images, fake audio, or political spam. The practical takeaway is simple: election-related AI tools will be judged not only by what they can generate, but by how well they reduce abuse.

Source: openai

Thursday, May 28: Google Brings Gemini Deeper Into Smart Homes

Google started rolling out a Gemini-powered Google Home feature that lets camera activity trigger smart home automations. For example, users can describe what they want a camera to notice, and that visual event can start a routine. The feature is limited for now, including U.S. English users in Google Home Public Preview with supported cameras and a premium subscription.

This is useful because it shows AI moving beyond chat windows. Instead of only answering questions, AI is beginning to understand visual events and take actions in the home. Still, Google says it should not be used for urgent safety or security situations, which shows this technology is promising but not fully dependable yet.

Source: theverge

Thursday, May 28: Microsoft Reportedly Prepares Its Own AI Coding Model

Reuters reported that Microsoft plans to unveil a new set of in-house AI models, including a coding model for GitHub Copilot, at its Build developer conference. The report also mentioned models for transcription, reasoning, speech, and images.

If confirmed, this would be a meaningful shift. Microsoft has relied heavily on partners such as OpenAI, but building more of its own models could give the company better control over cost, performance, and product direction. For developers, the most important thing to watch is whether Copilot becomes faster, cheaper, or more capable because of these models.

Source: reuters

Thursday, May 28: CNN Sues Perplexity Over AI Search Content

CNN filed a lawsuit against Perplexity, accusing the AI search company of unlawfully using and distributing CNN content. Perplexity responded that facts cannot be copyrighted.

This case matters because AI search tools sit directly between readers and publishers. If users get full answers from AI tools without visiting original articles, publishers worry that their traffic and revenue will fall. The result of cases like this could shape how AI companies use journalism, cite sources, and pay for content.

Source: reuters

Thursday, May 28: Mistral Defends Military AI and Expands Compute

French AI company Mistral defended the use of AI in military contexts and announced plans for a new data center in France. The company is positioning itself as a European AI alternative to U.S. tech giants.

The bigger story is AI sovereignty. Europe does not want to depend fully on American or Chinese AI systems, especially for sensitive areas like defense, government, and critical infrastructure. But military AI also raises serious ethical questions, so this area will remain closely watched.

Source: reuters

Friday, May 29: Samsung Ships Faster AI Memory Samples

Samsung said it started shipping samples of its 12-layer HBM4E memory chip to customers. HBM, or high-bandwidth memory, helps AI chips move large amounts of data quickly, which is important for training and running advanced AI models.

This matters because AI progress is not only about models. It also depends on chips, memory, servers, and data centers. Faster memory can help AI systems run more efficiently, and Samsung is trying to regain ground against rivals in the AI memory market.

Source: reuters

Friday, May 29: Wipro and ServiceNow Expand Agentic AI Workflows

India’s Wipro rose after expanding its partnership with ServiceNow to deploy agentic AI workflows across IT, HR, procurement, and cybersecurity. Agentic AI means systems that can take steps across a workflow, not just answer a single question.

For businesses, this is one of the most practical AI trends to watch. The goal is not just to generate text, but to reduce manual coordination inside companies. The challenge will be proving that these agents are reliable, secure, and worth the cost.

Source: reuters

Conclusion

This week showed three clear AI trends. First, coding agents are becoming a major battleground. Second, AI companies and publishers are still fighting over content rights. Third, the AI infrastructure race is getting bigger, from memory chips to data centers.

The main thing to watch next is whether these tools move from impressive demos to dependable daily use. That will decide which AI updates are truly useful and which ones are just noise.

Vijay Chauhan

Vijay Chauhan is a pro vibe coder with a passion for AI development and innovation. With deep expertise in crafting smart tools, he knows how to make AI dance to the rhythm of natural language. Always eager to share knowledge, Vijay blends tech mastery with creativity to build next-gen AI experiences.