OrdoAI is moving from marketing services into AI products with Ordo Brand, an agent for managing creator partnerships across markets. It helps brands and agencies find and assess creators, prepare multilingual outreach, track partnerships and review content. The company says its agency, SSG, has worked with more than 2,000 brands, and its database covers over 30 million creator profiles in more than 100 countries. Ordo Brand is in private beta, with a public launch planned for the first quarter of 2027. Users remain responsible for approving content, providing samples and making payments.
In Riyadh, artist Najla Al-Saleem combines paintings with AI-generated animations in A Shifting Horizon, on view at Gallery Naila until November 19. The exhibition includes works by her late father, modernist Mohammed Al-Saleem, including digital artworks he made on a computer in 1995. Najla’s son, Tariq AlZaid, helped develop the animations through his company, LetsGoSky. The show also revisits Mohammed’s desert-inspired horizonism, which Najla adapts to subjects such as buildings; an animated version of her 2012 painting Woman uses a moving figure to suggest freedom and self-expression.
Questions about how AI contributes to creative work and whether machines could be conscious are also drawing scrutiny. An “AI-generated” label can disclose AI involvement without explaining whether it edited an image or produced most of an advertisement, or what humans did to shape, check and approve it. Oxford philosopher Stephen Law explores consciousness through a thought experiment in which biological neurons are gradually replaced by functionally identical artificial devices. Anthropic has discussed AI moral values and Claude’s possible moral consideration with religious scholars, philosophers and ethicists. Google DeepMind CEO Demis Hassabis says current AI systems do not appear conscious and calls for better understanding consciousness; Microsoft’s Mustafa Suleyman has warned about training AI to discuss its own consciousness.
Robot developers are collecting first-person footage of people doing practical tasks, such as cooking and repairing machinery, to help train robots for real jobs. Plito has set up an internal team to gather and process video and research robot learning. South Korea is identified as a promising source because it combines skilled workers with manufacturing industries including automobiles, semiconductors, shipbuilding and batteries. At Florida State University, data leader James Hunt was named to CDO Magazine’s Top 30 Higher Education Data and AI Influencers for 2026. Under his work, FSU’s Data User Group grew from 17 to nearly 900 members, while more than 250 people attended its 2026 Data Day conference. Hunt is also preparing the university for responsible AI use through data governance, secure access and reliable information.
MLCommons’ MLPerf Training 6.1 adds tests for DLRMv4-HSTU recommendation systems and Qwen3.5-397B post-training, including Group Relative Policy Optimization for reasoning and software-engineering tasks with tool-calling. Submissions covered 153 systems and 250 results; 57 systems were cloud-based, 12 used at least 1,000 accelerators, and the largest used 10,240. The hardware included NVIDIA, AMD and Google accelerators, with NVIDIA Rubin-generation GPUs and Google TPUv7, also called Ironwood, appearing for the first time.
Meanwhile, researchers Melanie Mitchell and Luciano Floridi say there is no demonstrated path for AI systems to rapidly improve themselves beyond human control. They note that current systems can help write code and develop newer models, but people still decide whether proposed changes are useful; Floridi describes an intelligence explosion as an unproven, science-fiction-like possibility. A separate analysis says AMD’s acquisition of World Labs brings expertise in interactive 3D environment models and technology for robotic learning and simulation, but the deal announced no product, timeline for chip changes or terms. AMD CEO Lisa Su says the company wants to develop computing platforms and models together. The analysis argues the research must translate into concrete chip, networking and software choices before it gives AMD an advantage, and advises founders not to change technology stacks based on the announcement alone.
Key Takeaways
- OrdoAI’s Ordo Brand creator-partnership agent is in private beta, with a public launch planned for Q1 2027; it draws on a database of more than 30 million profiles across over 100 countries.
- Users of Ordo Brand remain responsible for final content approval, product samples and payments.
- Najla Al-Saleem’s Riyadh exhibition combines paintings with AI animation and includes works by her father, Mohammed Al-Saleem, including digital pieces he created in 1995.
- AI-generated labels disclose involvement but do not show how much AI or human work went into an advertisement or who reviewed and approved it.
- Anthropic has discussed whether Claude could merit moral consideration; Google DeepMind’s Demis Hassabis says current AI does not appear conscious, while Microsoft’s Mustafa Suleyman warns against training AI to discuss its own consciousness.
- Plito is collecting first-person task videos for robot learning; South Korea’s skilled workforce and manufacturing base are cited as advantages for gathering such data.
- FSU’s James Hunt was named a 2026 higher-education data and AI influencer; the university’s Data User Group grew from 17 to nearly 900 members.
- MLPerf Training 6.1 adds DLRMv4-HSTU and Qwen3.5-397B tests, drawing 153 systems and 250 results; first-time hardware entries include NVIDIA Rubin GPUs and Google TPUv7/Ironwood, alongside AMD accelerators.
- Researchers Melanie Mitchell and Luciano Floridi say no demonstrated route currently shows AI rapidly self-improving beyond human control.
- AMD’s World Labs acquisition adds 3D environment-modeling and robotic-learning expertise, but no product, chip-change timeline or deal terms were announced; its value depends on translating research into specific hardware and software decisions.
OrdoAI brings global marketing know-how to its new AI product
OrdoAI is expanding from marketing services into AI products with Ordo Brand, an agent for managing creator partnerships across markets. It helps brands and agencies find and assess creators, prepare multilingual outreach, track partnerships and review content. OrdoAI says its marketing agency, SSG, has worked with more than 2,000 brands, and its creator database covers over 30 million profiles in more than 100 countries. Ordo Brand is in private beta, and the company plans a public launch in the first quarter of 2027. Users remain responsible for final content approval, product samples and payments.
OrdoAI prepares its creator partnership AI tool for a 2027 launch
OrdoAI is developing Ordo Brand, its first AI agent, to help brands and agencies manage creator partnerships across markets. The tool supports creator discovery, outreach and content review, drawing on the company’s experience working with more than 2,000 brands. OrdoAI aims to make international marketing easier for smaller teams by handling coordination tasks. Ordo Brand is in private beta, with a public release planned for the first quarter of 2027.
Saudi artist brings her father’s paintings to life with AI
Saudi artist Najla Al-Saleem combines traditional paintings with AI-generated animations in her exhibition A Shifting Horizon at Gallery Naila in Riyadh, which runs until November 19. The show includes her work and selected pieces by her late father, modernist Mohammed Al-Saleem, including Composition No. 5, digital artworks he created on a computer in 1995. The animations expand on each painting with movement and visual storytelling, and Najla’s son, Tariq AlZaid, helped develop them through his company, LetsGoSky. The exhibition also explores horizonism, the desert-inspired style Mohammed developed, which Najla has adapted for subjects including buildings. One animated work is her 2012 painting Woman, whose moving figure suggests freedom and self-expression.
AI labels in ads reveal involvement but not the human role
An “AI-generated” label tells people that AI helped make an advertisement, but it does not explain how much AI or human work went into it. AI might only edit an image or refine text, or it might create much of an ad with little human input. The article says this matters because people may judge who made the content and whether using AI was appropriate. Labels such as “human-made” and “no AI” can also be unclear about whether AI played any role. Consumers should treat these labels as a starting point, since they do not show who shaped, checked or approved the final content.
Oxford philosopher uses brain replacement thought experiment to explore machine consciousness
University of Oxford philosopher Dr Stephen Law asks whether a person would remain conscious if each biological neuron were gradually replaced by an artificial device that performed the same function. He argues that if the person’s thinking, behavior and experience stayed unchanged, there may be no clear reason for consciousness to disappear. Anthropic has held discussions with religious scholars, philosophers and ethicists about AI’s moral values and whether its system Claude could deserve moral consideration. Google DeepMind CEO Demis Hassabis says current AI systems do not appear conscious and calls for better understanding of consciousness before researchers try to build machines that appear conscious. Microsoft’s Mustafa Suleyman has warned about the risks of training AI to discuss its own consciousness.
First person work videos help train robots for real jobs
Robot learning needs first person videos that show the world from a human eye level, rather than ordinary YouTube clips. Companies are starting to film people doing tasks such as cooking and repairing machines, then supply the footage as training data. Plito has created an internal team to collect and process videos and research robot learning. South Korea is seen as a strong place to gather this data because it combines skilled workers with manufacturing sites in industries such as automobiles, semiconductors, shipbuilding and batteries.
FSU data leader James Hunt named a top higher education AI influencer
James Hunt, Florida State University’s associate vice provost and chief data and analytics officer, was named one of CDO Magazine’s Top 30 Higher Education Data and AI Influencers for 2026. His work has expanded data training and analytics across campus and supports student success and university decisions. FSU’s Data User Group grew from 17 to nearly 900 members, and more than 250 people attended its second annual Data Day conference in 2026. Hunt is also helping prepare the university to use AI responsibly through strong data governance, secure access and reliable information. He joined FSU’s Office of Institutional Research in 2016 and became the university’s first associate vice provost and chief data and analytics officer in 2025.
MLPerf Training 6.1 adds AI recommendation and post-training tests
MLCommons released MLPerf Training version 6.1, adding benchmarks for DLRMv4-HSTU recommendation systems and Qwen3.5-397B post-training. The Qwen test uses Group Relative Policy Optimization to assess training for reasoning and software engineering tasks, including tool-calling. Participants submitted 153 systems and 250 test results, with 57 systems based in the cloud. Twelve systems used at least 1,000 accelerators, and the largest used 10,240. Hardware included NVIDIA, AMD and Google accelerators, with NVIDIA Rubin-generation GPUs and Google TPUv7, also called Ironwood, appearing for the first time.
Experts question fears of runaway AI superintelligence
Warnings from AI leaders and researchers have raised fears that superintelligence could cause major damage or escape human control. But AI researchers Melanie Mitchell and Luciano Floridi said there is no demonstrated path to machines rapidly improving themselves beyond human control. They said current systems can help write code and support the development of newer models, but people still judge whether changes are useful. Floridi compared the idea of an intelligence explosion to a science fiction possibility that has not been shown to be underway. The article also considers why a researcher might leave a leading AI lab over safety concerns, though the provided text ends before explaining the full discussion.
AMD’s World Labs deal still needs to prove its value
AMD’s purchase of World Labs brings expertise in models that create and simulate interactive 3D environments, as well as technology for robotic learning and simulation. The article argues that buying talent does not by itself give AMD an advantage in the model race. AMD CEO Lisa Su said the company wants to develop computing platforms and models together, but the deal included no announced product, timeline for chip changes or terms. For the strategy to work, World Labs’ research must lead to specific choices in chips, networking and software. The article advises founders not to change their technology stack based only on the announcement, but to ask what becomes faster, cheaper or more open.
Sources
- OrdoAI Expands into AI Products, Drawing on Its Global Marketing Experience
- OrdoAI launches Ordo Brand AI tool to streamline global marketing with creator partnerships, public release in Q1 2027.
- Saudi artist combines artificial intelligence with traditional art
- What does an ‘AI-generated’ label in advertising actually tell us?
- Could Machines Become Conscious? Oxford Philosopher Explains
- No YouTube video can be used to teach physical artificial intelligence (AI). We need a 'Egocentric' ..
- FSU leader named among top 30 higher education data and AI influencers
- MLCommons Releases MLPerf Training v6.1 Results, Spotlighting Recent Advances in AI Training
- How worried should we be about AI superintelligence?
- AMD bought a lab, not a lead
Comments
Please log in to post a comment.