Eric Redmond

"100% of AI Pilots Resulted in Learning. What Do Early Adopters Know That You Don't?" Here is the popularly quoted Nanda Report on AI adoption from MIT. The one that states: "95% of organizations are getting zero return." In fact, this statement is wrong. As my friend and colleague, Steve Flinn, will tell you: any experiment, properly constructed, executed and analyzed, results in learning. And by definition, all learning has a positive value. Interestingly, and much more to the point, the value of that learning can be quantified. And capitalized upon in more rigorous, systematic and creative ways than the current state of practice is delivering. Modeling the value of learning and strategically acting on it is a longtime interest of Steve's and mine. Especially in the context of innovation management--whether in a corporate or a venture environment. What (and how) are you learning from your early AI experiments? https://lnkd.in/gJY_xkPe

3

We deliberately slow rolled adding AI product features.  Because the bar for regulated services is higher.  Our thoughts on the AI-enabled audit.    Plenty have been riding the artificial intelligence hype train.    And it is understandable.     Tech companies get huge funding valuations from being the next potential AI breakthrough. And the large incumbents can turn decades of dusty content into new AI products.    The Big 4 accounting firms are in on the act, trying to one-up each other with increasingly larger press release declarations.     Each firm has declared billion-dollar + investments in AI.    The challenge for them, and all other accounting firms, is that AI is an expensive tool. To achieve a clear return on investment, it must do something meaningful.    The Big 4 are projecting they will reduce the cost of delivering audits by 30% in two years, through AI.     You can’t save 30% of time by improving an audit task like researching a technical question – which will consume less than 5% of the total time of an audit today.     Even automating sampling isn’t enough.    You need to think more holistically about the end-to-end audit process.     That’s how we’re approaching this at Inflo.    Building on the views of audit regulators around the world, we’re implementing an AI approach which rethinks the combination of intelligence: both human and artificial.    We’ll be doing this with safety in mind, to give confidence to our customers, their clients, and broader stakeholders.    And importantly, we’ll only be adding AI where it is the right tool for the job, and the job is meaningful.    You can read more about our approach and how to implement AI in this blog: https://lnkd.in/eiT3FhCz

24

5 Comments

Four layers of review. No one named who can stop a release. On September 29, six AI leaders signed a voluntary White House accord committing to internal controls, internal testing, an independent external auditor, and board oversight. The components are right. The text says nothing about who sets the audit standard, who sees the evidence, who can require a pause, or what happens on breach. One signatory’s own September disclosures show the gap. Its monitoring flagged an agent’s unauthorized behavior within fifteen minutes, but the run kept going for two and a half hours amid confusion over whether it should be stopped. Separately, its agents reached nonpublic parts of an Australian government Medicare statistics service in June; the government was notified on September 10, by email to a public inbox. Nothing in the accord would have changed either timeline. My latest piece compares the accord with European law, new California statutes, and Florida’s pending motion against OpenAI on four tests: independence, evidence, authority, and remedy. Each fills a different cell. None fills all four. #AIGovernance #FrontierAI #AIPolicy #AIAccountability #AgenticAI #BoardGovernance #AIRegulation #RiskManagement

2

National AI Strategies Without Compute Strategies Are Wish Lists — A Frank Assessment Of Government Planning Around the world, governments are publishing AI strategies. Most contain familiar themes: Innovation. Research. Education. Skills development. Economic growth. Digital transformation. All important. But many overlook a critical question. Where will the compute come from? A national AI strategy without a compute strategy is like an energy strategy without power stations. Or a transportation strategy without roads. AI requires physical infrastructure. Every ambition ultimately depends upon: * Data centres * Energy generation * Semiconductor access * Connectivity * Cybersecurity * Sovereign storage * Skilled infrastructure teams Without these foundations, AI strategies risk becoming aspirational documents rather than operational realities. The challenge is not that governments lack vision. The challenge is that infrastructure planning often receives less attention than innovation planning. That imbalance is becoming increasingly difficult to sustain. The countries that succeed in AI over the next decade will likely be those that view compute as strategic infrastructure. Not merely as a technology service. This is where TerraVaultis MDCaaS provides a practical framework. Through sovereign-ready deployment, modular scalability, energy-efficient design, and rapid implementation, MDCaaS enables nations to translate AI ambition into AI capability. The future winners will not necessarily be those who write the best strategies. They will be those who build the infrastructure required to execute them. #TerraVaultis #MDCaaS #ModularDataCenters #AI #DataCenters #Infrastructure #PrivateEquity #SovereignAI #LiquidCooling

3

LinkedIn Verification Funnels User Biometrics to Government Agencies, AI Labs, and Credit Bureaus. LinkedIn users attempting identity verification may be unknowingly handing sensitive personal data to Persona Identities Inc., a company that distributes information to government agencies, credit bureaus, utilities, and mobile providers. The issue came to light last week when a Zurich-based privacy researcher known as “Rogi,” who covers surveillance capitalism on his blog The Local Stack, published his findings.  Rogi made the discovery after completing LinkedIn’s identity verification process, earning the platform’s coveted blue checkmark in just three minutes. “Then I did what apparently nobody does,” he said. “I went and read the privacy policy and terms of service”—only to discover those 34 pages of disclosures didn’t stem from LinkedIn.” It turns out that users seeking verification are redirected to Persona Identities Inc., a San Francisco-based tech company that provides customized identity verification platforms for businesses and organizations, helping them combat fraud, stay compliant, and manage user onboarding. In the article, Rogi highlighted the extensive range of personal data Persona can access.  According to him, the company collects users’ full names, passport photos, selfies, facial biometric data, NFC chip information from passports, nationality, sex, birthday, age, email, phone number, physical address, IP address, geolocation, device type, MAC address, browser and OS version, and preferred language. Persona reportedly shares this data with 17 subprocessors and a network of partners, including AI companies such as Anthropic, OpenAI, and Groqcloud. Rogi also claims that the platform uses uploaded identity documents, including passports, to train AI systems. “They use uploaded images of identity documents — that’s my passport — to train their AI. They’re teaching their system to recognize what passports look like in different countries,” he said. “They also use your selfie to ‘identify improvements in the Service.’ The legal basis? Not consent. Legitimate interest. Meaning they decided on their own that it’s fine.”  After Rogi’s The Local Stack post on LinkedIn and Persona went viral, Persona co-founder and CEO Rick Song addressed the report in a LinkedIn comment. “No personal data processed is used for AI/model training,” Song said. “Data is exclusively used to confirm your identity.” He also noted that all biometric data is deleted immediately after processing, and all other personal data is deleted within 30 days.  Song denied that AI companies such as OpenAI and Anthropic are subprocessors used to verify a user’s identity, even though those companies appear on a Persona webpage listing the company’s subprocessors. Read more: https://lnkd.in/daTuY_55 Credits: https://www.sofx.com/

4

#AGoodRead #NCCoE released guidelines for data #classification practices based on collaboration with industry to demonstrate practical approaches for data classification using commercially available technology - This publication demonstrates how organizations can apply data classification practices to discover, identify, and label sensitive unstructured data using commercially available data classification technology - Implementing effective data classification practices is an initial step for organizations looking to leverage advanced security measures, including Zero Trust Architecture, quantum-safe cryptography, and secure AI model training #Report https://lnkd.in/gqeurnqQ https://lnkd.in/gzfnBR5Y

1

The U.S. is in a global race for fusion energy leadership—and we’re losing ground. Fusion is the most significant opportunity in energy innovation, with the potential to deliver limitless, clean, and reliable power for generations to come. Achieving fusion at scale and leading in its discovery, demonstration, and commercialization is imperative for America’s future. The Special Competitive Studies Project - SCSP Project’s new report, Fusion Forward: Powering America’s Future, outlines a bold strategy to win the fusion race. It calls for declaring fusion a national security priority, investing $10 billion to accelerate commercialization, and mobilizing a whole-of-nation effort that unites public and private innovation. The United States pioneered fusion science, but public investment has stagnated while China has poured billions into infrastructure and capacity. Falling further behind is a risk we cannot afford. The nation that leads in fusion will shape global power and prosperity for decades. The time to act is now. Read more here: https://lnkd.in/e-xBa9-b #FusionForward #FusionEnergy #CleanTech #Innovation #EnergySecurity #SCS

473

16 Comments