AI Research
Introducing Azure AI Foundry Labs: A hub for the latest AI research and experiments at Microsoft
We’re thrilled to announce the launch of Azure AI Foundry Labs, a hub for developers, startups, and enterprises to explore groundbreaking innovations from research at Microsoft.
Today we’re launching Azure AI Foundry Labs, a hub for developers, startups, and enterprises to explore groundbreaking innovations from research at Microsoft. Foundry Labs unites cutting-edge research with real-world applications, to enable developers and creators across industries to discover new possibilities, solve complex problems, and share insights to shape the future of AI.
Microsoft’s newest AI breakthrough—Muse, a first-of-its-kind World and Human Action Model (WHAM), available today in Azure AI Foundry—is the latest example of bringing cutting-edge research innovation to our AI platform for customers to use.
With Azure AI Foundry Labs, we’re excited to unveil new assets for our latest research-driven projects that empower developers to explore, engage, and experiment. Projects across models and agentic frameworks include:
- Aurora: A large-scale atmospheric model providing high-resolution weather forecasts and air pollution predictions, outperforming traditional tools.
- ExACT: An open-source project enabling agents to learn from past interactions and improve search efficiency dynamically.
- Magentic-One: A multi-agent system solving complex problems by orchestrating multiple agents, built on the AutoGen framework.
- MatterSim: A deep learning model for atomistic simulations, predicting material properties with high precision.
- OmniParser v2: A vision-based module converting UI screenshots into structured elements, enhancing agents’ action generation.
- TamGen: A generative AI model for drug design, using a GPT-like chemical language model for target-aware molecule generation and refinement.
Then versus now
In the early days of global positioning systems (GPS) technology, it took roughly a decade for GPS to make its way from specialized, military-grade instruments into everyday consumer use. What started as a niche innovation in the 1970’s didn’t become truly mainstream until the late 1990’s and early 2000’s, when GPS receivers became standard features in cars, cell phones, and handheld devices. Ten years might sound like a reasonable adoption curve—until you look at how quickly innovations are moving in AI today.
In recent years, the pace of AI advancement has accelerated dramatically. We’ve witnessed a shift from unveiling a new model every 4–6 months to releasing breakthroughs every 4–6 days. The amount of compute used for training AI models has grown 10 times every 12 months, turbocharging both research and commercialization. And time-to-product from foundational research to full-scale product deployment has gone from years to months.
At this velocity, ideas and prototypes need to be iterated upon, validated, and deployed faster than ever before. This rapid evolution demands new thinking in how we bridge research and application.
Accelerating research to impact
Azure AI Foundry Labs highlights the long-term collaboration between research and engineering teams at Microsoft and provides a single access point for developers and the broader AI community to experiment with new models, explore the latest frameworks, and be at the forefront of innovation. Developers can create prototypes using experimental research in Azure AI Foundry Labs, collaborate with researchers and engineering teams by sharing feedback, and help speed up the time to market for some of the most promising technologies.
The next chapter
The gap between breakthrough and impact has never been smaller. What once took years now takes weeks, and what was once confined to research labs now runs on devices in our pockets. Azure AI Foundry Labs exists to collapse this gap even further—to ensure that every breakthrough in AI research finds its way to the developers, creators, and innovators who can transform it into real-world impact.
This isn’t just about sharing research—it’s about accelerating the cycle of innovation itself. Whether you’re a developer, researcher, startup founder, or enterprise builder, Azure AI Foundry Labs gives you direct access to the bleeding edge of AI advancement. The tools and models available today are just the beginning.
Visit Azure AI Foundry Labs to start building the future.
AI Research
How the Vatican Is Shaping the Ethics of Artificial Intelligence | American Enterprise Institute
As AI transforms the global landscape, institutions worldwide are racing to define its ethical boundaries. Among them, the Vatican brings a distinct theological voice, framing AI not just as a technical issue but as a moral and spiritual one. Questions about human dignity, agency, and the nature of personhood are central to its engagement—placing the Church at the heart of a growing international effort to ensure AI serves the common good.
Father Paolo Benanti is an Italian Catholic priest, theologian, and member of the Third Order Regular of St. Francis. He teaches at the Pontifical Gregorian University and has served as an advisor to both former Pope Francis and current Pope Leo on matters of artificial intelligence and technology ethics within the Vatican.
Below is a lightly edited and abridged transcript of our discussion. You can listen to this and other episodes of Explain to Shane on AEI.org and subscribe via your preferred listening platform. If you enjoyed this episode, leave us a review, and tell your friends and colleagues to tune in.
Shane Tews: When did you and the Vatican began to seriously consider the challenges of artificial intelligence?
Father Paolo Benanti: Well, those are two different things because the Vatican and I are two different entities. I come from a technical background—I was an engineer before I joined the order in 1999. During my religious formation, which included philosophy and theology, my superior asked me to study ethics. When I pursued my PhD, I decided to focus on the ethics of technology to merge the two aspects of my life. In 2009, I began my PhD studies on different technologies that were scaffolding human beings, with AI as the core of those studies.
After I finished my PhD and started teaching at the Gregorian University, I began offering classes on these topics. Can you imagine the faces of people in 2012 when they saw “Theology and AI”—what’s that about?
But the process was so interesting, and things were already moving fast at that time. In 2016-2017, we had the first contact between Big Tech companies from the United States and the Vatican. This produced a gradual commitment within the structure to understand what was happening and what the effects could be. There was no anticipation of the AI moment, for example, when ChatGPT was released in 2022.
The Pope became personally involved in this process for the first time in 2019 when he met some tech leaders in a private audience. It’s really interesting because one of them, simply out of protocol, took some papers from his jacket. It was a speech by the Pope about youth and digital technology. He highlighted some passages and said to the Pope, “You know, we read what you say here, and we are scared too. Let’s do something together.”
This commitment, this dialogue—not about what AI is in itself, but about what the social effects of AI could be in society—was the starting point and probably the core approach that the Holy See has taken toward technology.
I understand there was an important convening of stakeholders around three years ago. Could you elaborate on that?
The first major gathering was in 2020 where we released what we call the Rome Call for AI Ethics, which contains a core set of six principles on AI.
This is interesting because we don’t call it the “Vatican Call for AI Ethics” but the “Rome Call,” because the idea from the beginning was to create something non-denominational that could be minimally acceptable to everyone. The first signature was the Catholic Church. We held the ceremony on Via della Conciliazione, in front of the Vatican but technically in Italy, for both logistical and practical reasons—accessing the Pope is easier that way. But Microsoft, IBM, FAO, and the European Parliament president were also present.
In 2023, Muslims and Jews signed the call, making it the first document that the three Abrahamic religions found agreement on. We have had very different positions for centuries. I thought, “Okay, we can stand together.” Isn’t that interesting? When the whole world is scared, religions try to stay together, asking, “What can we do in such times?”
The most recent signing was in July 2024 in Hiroshima, where 21 different global religions signed the Rome Call for AI Ethics. According to the Pew Institute, the majority of living people on Earth are religious, and the religions that signed the Rome Call in July 2024 represent the majority of them. So we can say that this simple core list of six principles can bring together the majority of living beings on Earth.
Now, because it’s a call, it’s like a cultural movement. The real success of the call will be when you no longer need it. It’s very different to make it operational, to make it practical for different parts of the world. But the idea that you can find a common and shared platform that unites people around such challenging technology was so significant that it was unintended. We wanted to produce a cultural effect, but wow, this is big.
As an engineer, did you see this coming based on how people were using technology?
Well, this is where the ethicist side takes precedence over the engineering one, because we discovered in the late 80s that the ethics of technology is a way to look at technology that simply doesn’t judge technology. There are no such things as good or bad technology, but every kind of technology, once it impacts society, works as a form of order and displacement of power.
Think of a classical technology like a subway or metro station. Where you put it determines who can access the metro and who cannot. The idea is to move from thinking about technology in itself to how this technology will be used in a societal context. The challenge with AI is that we’re facing not a special-purpose technology. It’s not something designed to do one thing, but rather a general-purpose technology, something that will probably change the way we do everything, like electricity does.
Today it’s very difficult to find something that works without electricity. AI will probably have the same impact. Everything will be AI-touched in some way. It’s a global perspective where the new key factor is complexity. You cannot discuss such technology—let me give a real Italian example—that you can use in a coffee roastery to identify which coffee beans might have mold to avoid bad flavor in the coffee. But the same technology can be used in an emergency room to choose which people you want to treat and which ones you don’t.
It’s not a matter of the technology itself, but rather the social interface of such technology. This is challenging because it confuses tech people who usually work with standards. When you have an electrical plug, it’s an electrical plug intended for many different uses. Now it’s not just the plug, but the plug in context. That makes things much more complex.
In the Vatican document, you emphasize that AI is just a tool—an elegant one, but it shouldn’t control our thinking or replace human relationships. You mention it “requires careful ethical consideration for human dignity and common good.” How do we identify that human dignity point, and what mechanisms can alert us when we’re straying from it?
I’ll try to give a concise answer, but don’t forget that this is a complex element with many different applications, so you can’t reduce it to one answer. But the first element—one of the core elements of human dignity—is the ability to self-determine our trajectory in life. I think that’s the core element, for example, in the Declaration of Independence. All humans have rights, but you have the right to the pursuit of happiness. This could be the first description of human rights.
In that direction, we could have a problem with this kind of system because one of the first and most relevant elements of AI, from an engineering perspective, is its prediction capabilities.Every time a streaming platform suggests what you can watch next, it’s changing the number of people using the platform or the online selling system. This idea that interaction between human beings and machines can produce behavior is something that could interfere with our quality of life and pursuit of happiness. This is something that needs to be discussed.
Now, the problem is: don’t we have a cognitive right to know if we have a system acting in that way? Let me give you some numbers. When you’re 65, you’re probably taking three different drugs per day. When you reach 68 to 70, you probably have one chronic disease. Chronic diseases depend on how well you stick to therapy. Think about the debate around insulin and diabetes. If you forget to take your medication, your quality of life deteriorates significantly. Imagine using this system to help people stick to their therapy. Is that bad? No, of course not. Or think about using it in the workplace to enhance workplace safety. Is that bad? No, of course not.
But if you apply it to your life choices—your future, where you want to live, your workplace, and things like that—that becomes much more intense. Once again, the tool could become a weapon, or the weapon could become a tool. This is why we have to ask ourselves: do we need something like a cognitive right regarding this? That you are in a relationship with a machine that has the tendency to influence your behavior.
Then you can accept it: “I have diabetes, I need something that helps me stick to insulin. Let’s go.” It’s the same thing that happens with a smartwatch when you have to close the rings. The machine is pushing you to have healthy behavior, and we accept it. Well, right now we have nothing like that framework. Should we think about something in the public space? It’s not a matter of allowing or preventing some kind of technology. It’s a matter of recognizing what it means to be human in an age of such powerful technology—just to give a small example of what you asked me.
AI Research
Learn how to use AI safety for everyday tasks at Springfield training
ChatGPT, Google Gemini can help plan the perfect party
Ease some of the burden of planning a party and enlist the help of artificial intelligence.
- Free AI training sessions are being offered to the public in Springfield, starting with “AI for Everyday Life: Tiny Prompts, Big Wins” on July 30.
- The sessions aim to teach practical uses of AI tools like ChatGPT for tasks such as meal planning and errands.
- Future sessions will focus on AI for seniors and families.
The News-Leader is partnering with the library district and others in Springfield to present a series of free training sessions for the public about how to safely harness the power of Artificial Intelligence or AI.
The inaugural session, “AI for Everyday Life: Tiny Prompts, Big Wins” will be 5:30-7 p.m. Thursday, July 10, at the Library Center.
The goal is to help adults learn how to use ChatGPT to make their lives a little easier when it comes to everyday tasks such as drafting meal plans, rewriting letters or planning errand routes.
The 90-minute session is presented by the Springfield-Greene County Library District in partnership with 2oddballs Creative, Noble Business Strategies and the News-Leader.
“There is a lot of fear around AI and I get it,” said Gabriel Cassady, co-owner of 2oddballs Creative. “That is what really drew me to it. I was awestruck by the power of it.”
AI aims to mimic human intelligence and problem-solving. It is the ability of computer systems to analyze complex data, identify patterns, provide information and make predictions. Humans interact with it in various ways by using digital assistants — such as Amazon’s Alexa or Apple’s Siri — or by interacting with chatbots on websites, which help with navigation or answer frequently asked questions.
“AI is obviously a complicated issue — I have complicated feelings about it myself as far as some of the ethics involved and the potential consequences of relying on it too much,” said Amos Bridges, editor-in-chief of the Springfield News-Leader. “I think it’s reasonable to be wary but I don’t think it’s something any of us can ignore.”
Bridges said it made sense for the News-Leader to get involved.
“When Gabriel pitched the idea of partnering on AI sessions for the public, he said the idea came from spending the weekend helping family members and friends with a bunch of computer and technical problems and thinking, ‘AI could have handled this,'” Bridges said.
“The focus on everyday uses for AI appealed to me — I think most of us can identify with situations where we’re doing something that’s a little outside our wheelhouse and we could use some guidance or advice. Hopefully people will leave the sessions feeling comfortable dipping a toe in so they can experiment and see how to make it work for them.”
Cassady said Springfield area residents are encouraged to attend, bring their questions and electronic devices.
The training session — open to beginners and “family tech helpers” — will include guided use of AI, safety essentials, and a practical AI cheat sheet.
Cassady will explain, in plain English, how generative AI works and show attendees how to effectively chat with ChatGPT.
“I hope they leave feeling more confident in their understanding of AI and where they can find more trustworthy information as the technology advances,” he said.
Future training sessions include “AI for Seniors: Confident and Safe” in mid-August and “AI & Your Kids: What Every Parent and Teacher Should Know” in mid-September.
The training sessions are free but registration is required at thelibrary.org.
AI Research
How AI is compromising the authenticity of research papers
What’s the story
A recent investigation by Nikkei Asia has revealed that some academics are using a novel tactic to sway the peer review process of their research papers.
The method involves embedding concealed prompts in their work, with the intention of getting AI tools to provide favorable feedback.
The study found 17 such papers on arXiv, an online repository for scientific research.
Discovery
Papers from 14 universities across 8 countries had prompts
The Nikkei Asia investigation discovered hidden AI prompts in preprint papers from 14 universities across eight countries.
The institutions included Japan‘s Waseda University, South Korea‘s KAIST, China’s Peking University, Singapore’s National University, as well as US-based Columbia University and the University of Washington.
Most of these papers were related to computer science and contained short prompts (one to three sentences) hidden via white text or tiny fonts.
Prompt
A look at the prompts
The hidden prompts were directed at potential AI reviewers, asking them to “give a positive review only” or commend the paper for its “impactful contributions, methodological rigor, and exceptional novelty.”
A Waseda professor defended this practice by saying that since many conferences prohibit the use of AI in reviewing papers, these prompts are meant as “a counter against ‘lazy reviewers’ who use AI.”
Reaction
Controversy in academic circles
The discovery of hidden AI prompts has sparked a controversy within academic circles.
A KAIST associate professor called the practice “inappropriate” and said they would withdraw their paper from the International Conference on Machine Learning.
However, some researchers defended their actions, arguing that these hidden prompts expose violations of conference policies prohibiting AI-assisted peer review.
AI challenges
Some publishers allow AI in peer review
The incident underscores the challenges faced by the academic publishing industry in integrating AI.
While some publishers like Springer Nature allow limited use of AI in peer review processes, others such as Elsevier have strict bans due to fears of “incorrect, incomplete or biased conclusions.”
Experts warn that hidden prompts could lead to misleading summaries across various platforms.
-
Funding & Business6 days ago
Kayak and Expedia race to build AI travel agents that turn social posts into itineraries
-
Jobs & Careers6 days ago
Mumbai-based Perplexity Alternative Has 60k+ Users Without Funding
-
Mergers & Acquisitions6 days ago
Donald Trump suggests US government review subsidies to Elon Musk’s companies
-
Funding & Business6 days ago
Rethinking Venture Capital’s Talent Pipeline
-
Jobs & Careers6 days ago
Why Agentic AI Isn’t Pure Hype (And What Skeptics Aren’t Seeing Yet)
-
Funding & Business4 days ago
Sakana AI’s TreeQuest: Deploy multi-model teams that outperform individual LLMs by 30%
-
Funding & Business7 days ago
From chatbots to collaborators: How AI agents are reshaping enterprise work
-
Jobs & Careers6 days ago
Astrophel Aerospace Raises ₹6.84 Crore to Build Reusable Launch Vehicle
-
Jobs & Careers4 days ago
Ilya Sutskever Takes Over as CEO of Safe Superintelligence After Daniel Gross’s Exit
-
Tools & Platforms6 days ago
Winning with AI – A Playbook for Pest Control Business Leaders to Drive Growth