Microsoft AI: What's the Microsoft Approach?
AI technologies and tools have been created with every organization and individual in mind, at any level of any hierarchy, in mind. AI has made its presence known in workplaces, homes, universities, manufacturing plants, research labs, and academic institutions all over the globe to farmers, software developers, scientists, salespeople, and security personnel.
We have made significant investments in AI as we believe it can benefit individuals, industries, and society. And we aim to bring technology and humans together responsibly to fulfill AI's promises.
Explore Microsoft AI to gain more knowledge.
- Our view on ai and its use responsibly
- AI research: our approach
- AI infrastructure: our approach
- AI and social benefit: our view
How to Use AI Responsibly
Microsoft believes strongly in developing and using powerful technologies responsibly, so our AI design adheres to five core principles of fairness, reliability, and safety; privacy/security protection; and inclusiveness/transparency/accountability.
Microsoft applies these principles across its entire company to develop and deploy AI that positively contributes to society. Natasha Crampton is Microsoft's Chief AI Officer; according to her, cutting-edge technologies can be safely introduced into society to make people more productive.
Artificial intelligence systems are created from many decisions taken by humans during design and implementation, which must all lead to positive and equitable results, from purpose to people's interactions.
Crampton emphasizes the importance of responsible AI by noting, "Our practice of responsible AI involves considering AI from every stage, including design and lifecycle evaluation. This ensures that appropriate controls and mitigations can be built-in before being added as afterthoughts at any point during production."
It may not eliminate all risks, but this approach encourages developers to be clear about any limitations, consider potential uses or misuses for systems they develop, and think expansively about ways to secure its benefits while mitigating its risks. However, it prompts developers to consider all associated with any particular system, its risks, benefits, intended use/misuse considerations, etc.
Proactive self-regulation by companies is certainly one route forward; however, not every business will adopt these practices. Brad Smith of Microsoft highlighted this importance during an AI moment at work by calling upon policymakers to be thoughtful when approaching AI policy decisions.
Smith urges countries and communities to use democratic lawmaking processes to have an in-depth dialogue about setting boundaries that protect individuals under the law. Effective AI regulations must be outcome-focused and durable and prioritize high-risk applications to keep pace with rapidly advancing technologies and shifting social expectations. "To spread AI's benefits widely," according to one author, global regulatory approaches will have to be flexible enough and adaptable enough "just like AI itself.
Microsoft strongly believes in democratic lawmaking processes as an essential means to facilitate global dialogue among industry, academia, and civil society stakeholders about AI development and use. Their dialogue establishes principles and norms to enable organizations to develop AI solutions responsibly.
As part of our process, we are dedicated to publicly disseminating all learnings, best practices, and tools utilized. The responsible AI standard serves as a resource to translate our principles at an abstract level into practical guidance for engineers.
AI technology poses risks that must be anticipated and mitigated. Some may abuse AI to target individuals with false information or undermine democracy - these must all be managed together for our society's well-being.
Microsoft's responsible AI initiative has been an ongoing cross-company and cross-discipline effort for more than six years, providing us with a solid basis to build for tomorrow. Our vision remains clear - moving forward is necessary to keep pace with technology's rapid advancement and society's changing expectations.
Crampton states that our work in AI stems from its commitment to helping all organizations, regardless of size or sector. "AI should only ever be used responsibly by its creators or users."
AI Research: Our Approach
Microsoft research is an influential research organization with over three decades of experience advancing computing fundamentals. Translating scientific advances into innovative technologies that benefit customers and society.
Researchers from around the globe collaborate to develop cutting-edge technologies in speech recognition and translation, image captioning systems, understanding natural language usage, and answering everyday common sense queries. These researchers have made groundbreaking advances using AI for speech recognition, machine translation, and captioning systems; they even develop image captioning systems recognizing facial features as they come out recognizing faces for recognition!
Recently, efforts have centered around creating large-scale models capable of processing information more precisely while remaining easy for people to comprehend and use. Microsoft Azure's supercomputing and deep learning advances, combined with massive datasets from the internet, enable AI models to perform different functions across many applications.
Ashley Lorens is vice president and managing director at Microsoft Research. According to her, large-scale AI has fundamentally altered the computing landscape. As we adapt to this shift, new frontiers will open up that help us explore how human and machine bits of intelligence interact and reinforce each other in unexpected and profound ways.
Thanks to our research, we have developed Artificial Intelligence technologies that are more flexible and efficient. Microsoft Research has spent many years tackling these problems head-on and has developed expertise in areas like parallel computing - which enables users to train machine-learning models faster than ever and at unprecedented scales - and deep-learning optimization libraries like DeepSpeed that enable distributed training of models.
Microsoft products and services are tailored towards solving real-world issues. Project FarmVibes uses AI and data to provide open-source tools that aid farmers in adopting sustainable farming practices. In contrast, Microsoft's climate research Initiative provides research resources and computing power for multidisciplinary scientists working against climate change.
AI4Science is dedicated to applying deep learning in natural sciences to accurately predict and model natural events, helping address climate change and green energy concerns. Project teams comprising our global researchers and engineers are exploring an innovative machine-learning method by using fundamental equations instead of online datasets as training data sources.
Llorens notes that such research could enable scientists to better comprehend and predict natural phenomena on all scales ranging from quantum to galactic levels, leading to breakthroughs such as discovering materials capable of extracting carbon from the atmosphere.
He states, "Our ultimate objective is to expand technological frontiers that enhance human experiences while benefitting society." To achieve our aim of progress and innovation, we must join as part of an international research community that challenges preconceptions about what's possible and helps shape a roadmap to future developments.
Related:- Microsoft Azure Development - Build SaaS, PaaS and IaaS
AI Infrastructure: Our Approach
Customers want a partner that will assist them in adopting AI more rapidly and efficiently across all workloads to increase efficiency, enhance operations, and decrease costs. We predicted the rapid demand surge for AI-powered systems over a decade prior. We began planning accordingly by creating a special computing infrastructure.
Microsoft Azure AI provides infrastructure designed and optimized to run large AI models efficiently and increase productivity and creativity. Through our investments, we're providing clients with AI products tailored specifically for them, as well as providing optimal performance and scale when inference or training workloads arise.
Our cutting-edge architecture features fast GPUs and an innovative networking infrastructure connecting thousands of GPUs for AI model inference and training at scale. Eric Boyd is Microsoft's corporate vice-president for the AI Platform. He believes having many GPUs connected through high bandwidth connections enables everything else.
Organizations of all sizes are turning to Azure AI because it helps them achieve more results faster, with greater ease, thanks to enterprise-grade security, privacy, and AI protections provided by Azure.
Azure will serve as our AI supercomputer of the future and serve our mission to democratize AI. Microsoft is pushing cloud supercomputing technology forward; their first supercomputer ranked amongst the top five by 2023, while multiple AI supercomputing systems were constructed massively at scale.
Working closely with OpenAI, we enhanced our AI infrastructure to deploy their models and train them for researcher advancement. This service is now available to all Azure customers.
Boyd stated when other parties come to us seeking OpenAI-like technologies, "we will provide them with our infrastructure as it stands today." Microsoft Azure OpenAI offers developers and businesses high-performing AI models with industry-leading uptimes - such as GPT-3.5 and Codex models used by its products, such as GitHub Copilot and power platform. These same models also help drive uptime at these products from Microsoft.
Our AI infrastructure constantly adapts based on feedback from customers and partners when training or serving AI models. Working closely with industry partners, we design GPUs, data centers, and networks specifically optimized for AI workloads. Boyd asserts that Microsoft remains at the cutting edge of technology, offering its customers unparalleled training, software, and networking services infrastructure.
Our Approach to Using AI for Social Good
Microsoft believes AI holds great promise to address some of humanity's most urgent issues. Their AI for good initiative equips individuals and non-profits with funding, expertise, and technology needed to expedite progress on issues like access, digital literacy, equity, climate change sustainability, human rights, food security, cybersecurity, and beyond.
Juan Lavista Ferres, Microsoft's Chief Data Scientist and AI for Good Lab Director. According to him, artificial intelligence can offer solutions that may not be optimal yet are necessary in certain instances. He asserts that AI may provide solutions that are not necessarily better but at least viable enough.
One such project entails mapping peoples' locations to help understand disaster risks and direct preparation efforts more accurately. One such mapping project in India involved cataloging building roof materials used, which allowed disaster teams to prioritize those most likely to collapse during specific natural disasters. Another initiative involved mapping renewable energy installations worldwide to demonstrate their environmental impact.
Lavista Ferres notes that, while humans could spend up to 400 years studying satellite images and determining whether humans were present, AI models can complete this task within one hour, allowing people to quickly gather all relevant information at one location to make decisions.
Microsoft's AI for Good Lab is an applied research lab dedicated to data visualization and applied research that leverages Azure cloud computing and big data at its peak potential. Their team of data scientists collaborates closely with experts and strategic partners from universities, non-profits, and governments worldwide in addressing global issues and pinpointing areas where assistance might be helpful.
Microsoft and our partners collaborate on numerous projects together, with one being our AI for Health Program that invests in research from institutions like Fred Hutchinson Cancer Research Center, IRIS Novartis Foundation Seattle Children's Research Institute to make healthcare affordable to everyone, particularly where there are less doctors but greater medical need.
Microsoft Reveals a 'Wave 2 Release Roadmap' for Dynamics 365
Microsoft revealed its Wave 2 product release plans this week, including its Dynamics 365 enterprise resource planning solutions and power platforms (Power BI, Power Apps, and Flow).
Microsoft unveiled three AI capabilities as part of its Wave 2 plan at the business applications summit, and these should become available this year or next. However, their announcement was somewhat overshadowed by that of Salesforce.com who revealed plans to acquire Tableau software--an analytics software provider--for $15,7 billion worth of Salesforce shares and to use.
Microsoft Dynamics 365 combined with the power platform is widely recognized for using data analytics to gain data insights. According to research in a blog posting, business intelligence markets are consolidating. He suggested companies could protect themselves by employing new BI fabrics that streamline management issues across various BI platforms.
Microsoft's wave 2 announcements for dynamics 365 & power platform are an open and transparent presentation of product plans; exact dates were often not given when discussing new features or upgrades to existing functionality.
Wave 2 features are expected to arrive between October 2023 and March 2023 and could become available as early as August 2, the general availability date has been scheduled for October 1, Since last year, Dynamics 365 has followed an iterative two-year feature update cycle where new features are typically released during spring and autumn releases.
AI Build for PowerApps and Flow
Microsoft unveiled its Wave 2 plan on June 10, focusing on adding AI enhancements to power platform products such as powerapps & Flow. AI Builders were unveiled for both products; PowerApps is tailored for business users creating apps directly, while Flow is used for automating tasks by connecting workflows and processes.
AI builder introduces "common AI situations" to app developers. According to an announcement, this can solve daily tasks like form processing, object recognition, text and binary classifiers, and classification. AI Builder works with the flow to pull data and execute actions; Microsoft suggests using this for credit checks as one use-case scenario.
Microsoft also touts its ability to develop "attestable workflows" using power platform tools combined with Azure blockchain service to create "attestable workflows." Azure blockchain service serves as Microsoft's electronic ledger storage solution. Also, it offers assurances when working on supply chains with multiple parties involved.
PowerApps Portals
Microsoft announced the new PowerApps Portals feature will soon become a public preview and allow organizations to build websites accessible by other members. At the same time, external visitors may sign into them using various identity providers like LinkedIn or personal accounts and Azure Active Directory organizations. PowerApps portals can be created through customizable templates. According to Microsoft, such websites were only previously available as an add-on within Dynamics 365.
PowerBI
Microsoft Power BI, its business intelligence software solution, continues to gain new AI features through Azure Cognitive Services and Machine Learning. New text analysis features, image recognition capabilities, and handwriting analysis will be integrated. Visual analytics powered by AI include Decomposition Tree AI and Distribution Changes. At the same time, models created within Power BI will also be exported to Azure machine learning for further enhancement. Microsoft has pledged that models created within Power BI will also be exportable into Azure machine learning for further training.
Microsoft claims the Power BI Q&A function can be trained to recognize and adapt company-specific terminology like synonyms, phrases, or domains. Microsoft will update Power BI to resemble other Office 365 apps more closely and offer users collaborative features such as report comments.
Announcements
Microsoft announced additional announcements beyond its Wave 2 releases: Power BI's paginated reports are now publicly available, reaching "general capability," meaning they are ready for commercial usage. Paginated reports, powered by SQL server reporting services, enable Power BI users "to quickly create and distribute highly formatted, pixel-perfect reports alongside their interactive Power BI content. Reports can then be archived or printed as Word or PDF documents." Microsoft stated that data could be pulled from Power BI, Azure analysis services, and other database management systems such as SQL server, Oracle, and Teradata. Paginated reports are exclusive to organizations with an active Power BI premium subscription plan.
Power BI template apps have now reached "general capability." Microsoft defines Template Apps as anything from website analyses and analysis tools for social media to complex data modeling solutions that target key business parameters. Template Apps will soon replace Content Packs as the solution of choice.
Users with iOS 12 devices or higher now have Siri voice shortcuts as part of the Power BI mobile app preview phase on iOS devices running iOS 12, 11 or greater. Spotlight searches for Power BI are available on these iOS devices. This week, Microsoft also provided details regarding Power BI's June and flow updates that occurred over May and into June.
Microsoft Dynamics 365 Artificial Intelligence: Features
Microsoft Dynamics 365 Artificial Intelligence allows your team to gain insight, enhance customer experiences and make better decisions through AI. AI features have been specifically created to assist users in breaking down data silos and creating machine-learning-powered experiences in sales, marketing and customer service - S.A.Global is your partner in Data and AI solutions.
- Customized transformation: Use AI as you see fit, using intuitive interfaces and detailed instructions with applications for an integrated, seamless experience.
- Eliminate siloed models of business: By unifying data across applications for an improved understanding of clients and your organization, integrate analysis and insight across touchpoints for increased relevance to client behaviors and needs.
- Your business's needs can be easily fulfilled: Customizable AI features enable rapid adaptation to changing market demands more swiftly and agilely than before.
Microsoft Announces Plans to Enhance AI Capabilities & Introduce Two New Dynamics 365 Applications
Microsoft recently unveiled that they will add advanced AI features and capabilities in 2023. Expanding AI across its products was welcome news, but even more surprising was Microsoft's plan to introduce two brand new Dynamics 365 applications: Finance Insights & Project Operations will join existing ones as soon as next quarter.
What is Dynamics 365 Finance Insights?
From May, Dynamics 365 finance insights will be available as a preview and provide your finance team with AI capabilities. AI can help gain better insight and understanding into financials and take actions to increase cash flow. Automation features allow more time to be dedicated to analyzing insights and moving your company forward.
What is Dynamics 365 Project Operations?
Starting October 1 this year, the Dynamics 365 Project Operations app will replace its predecessor, the Dynamics 365 Project Service Automation app, to facilitate better collaboration among project teams by merging Microsoft Project with Dynamics 365 Project Service Automation. Furthermore, this application offers greater visibility, collaborations, and insight for successful project completion.
Are you already utilizing dynamics 365 project service automation, or are you concerned about how this will impact your business? Microsoft will provide up to 4 years of coverage when migrating from project service automation to project operations.
Microsoft Also Disclosed What Else?
Microsoft recently unveiled changes it would make to Dynamics 365 applications. As part of these modifications, Customer Insights allows customers to enhance customer profiles by connecting to first- and third-party data sources; activation or update could occur immediately with such profiles activated or updated in real-time. Microsoft announced as part of Wave 1 it will expand Customer Insights availability across GCC environments to enhance citizen experiences and further optimize citizen satisfaction.
Dynamics 365 for sales and Insights has also recently unveiled new manual and predictive forecasting features. This allows sales teams to understand their pipelines better and predict more accurately. They will also have access to more comprehensive performance predictions in the future using data gleaned from your CRM, such as current leads, opportunities won/lost, opportunities won, and contacts/accounts. Dynamics 365 creates forecasts from this information, which accurately forecasts future performances using models built using these figures based on past trends. It is designed so even non-scientists can use/understand its use! This feature has been created so anyone can take full advantage of it no matter their background knowledge level or not being an expert user.
Want More Information About Our Services? Talk to Our Consultants!
What Are the Next Steps?
AI technologies as much as AI holds immense promise to address our current problems, its limitations must also be considered. Lavista Ferres stresses the need to consider moral and ethical concerns with each project to ensure models receive adequate data that doesn't leave anyone out.
"AI models generalize their models to fit only what data it already possesses; any bias could present problems." Human input should always remain paramount: We strive for solutions that benefit everyone equally and not just certain cohorts of users.
