Innovative Solutions’ Post

#AI technology and its social and regulatory context are changing rapidly. Organizations are grappling with how to provide the right oversight of their AI initiatives. Three things are necessary for good oversight: (1) a way to uncover or sense risks that might not have been present or evident before, (2) an understanding of the risks’ severity and the possibilities open for mitigating them, and (3) an ability to make rapid and effective risk-based decisions that balance risk with agility and innovation despite constant change and uncertainty. My colleague Mark Schwartz has crisply articulated these considerations. https://lnkd.in/gNyJHdJs

3 Ways to Succeed in the Rapidly Changing AI Landscape

Why #AI governance should be step 1 in your #GenAI journey. #ibmpartners #watsonx

It is a necessity for AI-driven companies to direct, manage and monitor the artificial intelligence (AI) activities. Let's walk the path of a trustworthy AI.

Kicking off 2024 is the growing interest among financial sponsors in the upside of artificial intelligence. It’s no surprise that private capital is flooding into a technology that has become the biggest investment priority for global businesses according to a recent survey by KPMG, but how sponsors are looking to seize that opportunity is perhaps less expected. We examine their desire to target proprietary data sets in order to create new businesses harnessing pre-trained AI models and explain the scope of the due diligence and some of the ways these tricky deals can be de-risked. #ai #privateequity #investmentopportunities #artificialintelligence #derisking

As businesses begin to embed generative AI, there is much to consider such as: ☑ Fairness ☑ Privacy ☑ Transparency Having an #AI governance strategy addresses these challenges. General Manager, IBM Ecosystem Kate Woolley discusses why AI governance is the first step of the #genai journey, and how to get started: https://lnkd.in/e3g2g7UT IBM watsonx IBM Data, AI & Automation #aiadoption #trustedai #ibmpartners #ai4good

🤖 Adopting AI in a responsible way 🤖 In a world brimming with AI possibilities, it's crucial to focus on practicality when beginning your AI journey. 🔍 Discover particle AI strategies 🔍 → Build a strong data foundation → Explore advanced analytics → Harness Generative AI and more → Address security & privacy concerns → Implement safeguards & governance → Empower, not replace, workers with Ai AI and advanced analytics hold transformative potential, but a practical, responsible approach ensures you reap the rewards. Dive into the details here: https://bit.ly/3UfSnrR Stay informed, stay ahead! For tailored advice and optimal strategies, reach out to our office. #AI #DataAnalytics #ResponsibleAI #DigitalTransformation

AI Governance + WatsonX.governance. If you know of any other company baking governance into AI, please share with me. Microsoft Co-Pilot? Hmm.. how good is your knowledge of Microsoft Purview? I recently had breakfast with a Microsoft exec who related how Co-pilot uses Purview for all its governance. Really? Have you see what is happening to Purview? How it is being split apart? Governance as a baked-on, afterthought, if you would care for my opinion. What about Bing (OpenAI ChatGTP) in the enterprise? Where is your governance there? We do all remember the initial rollout of Windows 10 and open warfare with Microsoft and those of us who did NOT want to share ALL of our data with Microsoft. Windows 10. Then Windows 11 requiring us (at least it made it hard to see a local login option) to create a MS ID. Why? to collect data on us. There is not a good record of Governance with Microsoft. IBM has set a standard and hope other's learn and adapt their AI's to stronger governance. We see the White House writing AI standards and seeking governance as well this spring. Businesses should not wait for the gov't - because then they will create a new agency to manage AI compliance - forcing companies to retool their AI- often at immense expense. What AI tools are you aware of with governance and what does that governance provide?

Isn’t it interesting how the more complex new technology gets, the more its development and deployment rely on multi-disciplinary teams, team members with diverse backgrounds, and engagement with broader societal actors? Launching new products isn’t just about knowing the customer anymore. It’s promising to see the many, varied approaches around the world to setting guardrails for artificial intelligence. This offers an excellent laboratory for countries and economic regions to test and learn from each other.  Values reflect the cultural context. I wonder, as companies and regulatory bodies around the world advance their ethical guardrails for AI, is it possible that we might see a convergence of global values as we learn from each other? Relatedly, check out “Bring Human Values to AI” in the latest HBR Mar-Apr 2024, by Jacob Abernathy, François Candelon, Theodoros Evgeniou, Abhishek Gupta and Yves Lostanlen. The authors emphasize that AI value alignment is not just a regulatory issue but also a product differentiator. This article offers concrete recommendations for companies along 6 Challenges in bringing AI-enabled products and services to market. 1.        Define Values for your Product ·      Embed established principles. Anthropic uses the United Nations’ Universal Declaration of Human Rights as the principles guiding its AI Assistant, Claude. ·      Articulate your own Values. Set up a diverse team of experts or enlist customers and employees in the process.  2.        Write the Values into the Program ·      There are different methods for constraining the behavior of AI systems such as privacy by design and safety by design and building in the necessary feedback loops.  ·      Generative AI will require more formal red lines (company + regulators) for adhering to defined values. ·      Because guardrails and regulations evolve over time, these must be embedded in the AI’s programming so that any changes apply across systems. ·      Tracking compliance with values includes tracking user’s conduct as with aggressive behavior in online gaming or terrorist material on social media platforms. 3.        Assess the Trade-Offs ·      Privacy vs accuracy ·      Time to market vs risk of values misalignment ·      Oversight Boards can be effective for weighing value-driven decisions 4.        Align your Partners’ Values ·      Companies must understand the values underlying the models and training data they use from partners. ·      Owners of foundational models have little control in how partners modify them for their own purposes. 5.        Ensure Human Feedback ·      Ensuring values alignment requires the processing of large volumes of data and human review. 6.        Prepare for Surprises ·      Companies must deploy a robust system for detecting and limiting harmful or unexpected behavior of it AI systems. ·      Participation in AI-incident databases (eg OECD or Partnership on AI) is recommended for continued learning.

Forget about the technology; let's first focus on AI governance as the foundation of trustworthy, generative AI! According to Business Insider, starting with humans first instead of technology is key. It's important to figure out your must-haves and can't-dos first. This will make selecting the right AI tools easier because you will know your boundaries, ethical requirements, and risks. Check out this article to learn more about why AI governance should be the first step in your generative AI journey. #AIgovernance #trustworthyAI #generativeAI #techethics