AI Impact: Why is trust the key to unlocking value with AI?

By Adam Mills and Gabriel Snyder

See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.

Newsweek

Editor's note: This is AI Impact, Newsweek's weekly newsletter where each week, we will explore how business leaders are unlocking real value through artificial intelligence.

Tap or click here to get this newsletter delivered to your inbox.

Inference Layer

Why Trust Is the Key to Unlocking Value with AI

By Gabriel Snyder

In Davos last week, I had the honor of moderating a dinner discussion hosted by Cognizant and Newsweek which convened CEOs, CTOs, chief AI Officers and transformation executives from sectors ranging from financial services and consumer goods to technology firms. The launching point was an op-ed by Cognizant CEO Ravi Kumar, published in Newsweek that laid out the central challenge of extracting value from AI: “Businesses struggle with an elongated J-curve in which initial investments and adoption realize modest returns… The fastest way to close this gap is to help businesses compress the J-curve by accelerating adoption in the enterprise.”

In the lively conversation (which was held under Chatham House Rules, so sorry, no direct quotes) one of the main organizing principles that emerged was the concept of trust. Now, trust is a frequently invoked word, but discussion surfaced how trust operates across fundamentally different dimensions depending on industry context. Financial services leaders frame trust as security infrastructure—the certainty that transactions execute safely and data remains protected. Consumer goods executives think of trust as product integrity—the assurance that what reaches customers meets quality and safety standards.

But these industry-specific manifestations point to a deeper substrate: trust is what allows any brand to maintain permission to operate in its chosen domain. Customers trust financial institutions to safeguard their money, consumer brands to deliver safe products, technology companies to protect their data. This permission structure—the implicit contract between enterprise and customer—defines the boundaries of what organizations can successfully deploy.

AI introduces a categorical shift in this trust architecture. Customers and employees haven't yet granted AI systems the permission they've extended to human-operated processes. This manifests as a management challenge today—organizations struggle to drive adoption of AI tools, encounter resistance to AI-generated outputs, face skepticism about algorithmic decisions.

My main takeaway from the discussion in Davos, though, is that trust is now the existential business question for the agentic AI economy. When AI systems make autonomous decisions, execute independent actions and represent organizations directly to customers, the permission structure becomes the primary constraint on value creation. The velocity gap isn't primarily technical. It's a trust deficit that determines which organizations can convert AI capacity into operating results.

Core Intelligence

Building with Care

DeepMind is one of the most influential AI research organizations in the world, and in a recent conversation with Marcus Weldon, Newsweek’s contributing editor for AI, the lab’s chief operating officer, Lila Ibrahim, described how difficult it is to pursue cutting-edge breakthroughs while maintaining deep responsibility.

At DeepMind, the quest isn’t simply to push capabilities forward. It is to do so responsibly, embedding care and ethical reflection into research practices rather than treating them as afterthoughts. As Ibrahim put it, the guiding question for her team is, “How do we make sure that the AI is happening with us and not to us?”—highlighting a commitment to agency, oversight and alignment that goes beyond performance benchmarks.

That responsibility mandate sits alongside an ambition to solve some of the hardest problems in science and computing. DeepMind’s most celebrated work—like the AlphaFold models that have revolutionized protein folding and earned a Nobel Prize—illustrates how far AI can stretch human understanding when powerful models are directed at meaningful scientific challenges. But those successes also underscore the difficulty of translating general intelligence research into systems that operate safely and predictably in the world.

The emphasis on pioneering responsibly changes what success looks like. It places human impact and societal context on par with technical innovation, shaping how research priorities are selected, evaluated and governed. That dual mandate—breakthrough and care—is not merely a slogan; it defines DeepMind’s operating rhythms, hiring outlook and internal deliberation.

For business leaders and policymakers, the takeaway is clear: frontier AI research isn’t just about raw capability gains. It’s about embedding human values into the design of transformative systems at the earliest stages of development. That approach doesn’t guarantee smooth paths to commercialization or short-term returns, but it may offer a model for organizations navigating the same tension between performance and responsibility.

In the emerging era of powerful AI, the real test of leadership may not be who builds the smartest models—but who builds them with intent, oversight and a sense of shared human stake.

You can read Weldon’s full analysis of the discussion here: Going deeper inside the mind of DeepMind: The thinking labs.

AI Impact Awards & Summit

The Newsweek AI Impact Awards seek to identify and recognize uniquely innovative AI solutions that solve critical business problems in different industry segments, or significantly advance capabilities. Recognition comes not from ideas, but from measurable IMPACT on business operations.

Prompt Injection

Chris Burchett | Senior Vice President, Generative AI, Blue Yonder

...

"One of my most significant recent realizations about AI is how dramatically large language models have advanced in both reliability and capability. Their improvements in code generation and agentic reasoning have raised the bar for what AI agents can accomplish with minimal human intervention.

When these more capable models are paired with structured decision-making frameworks such as the ‘SADA Loop,’ organizations benefit from agents that see changes in their environment, analyze the new information, decide on the best path forward, and then—and only then—act accordingly. The underlying model improvements alone have made each phase of this loop faster, more precise and more reliable.

This has reshaped my view of how organizations should think about AI adoption. The technology is no longer the limiting factor—deployment is. Agents are now capable of bringing machine-level speed and precision to proactively address complex enterprise challenges, but only if organizations actively implement and refine them.

There is ultimately no substitute for hands-on experience. The most successful organizations will be those investing the time to understand, configure and continually improve these systems alongside their developers."

Have your own lesson to share? Email us at: ai.newsletter@newsweek.com

Run Log

By Adam Mills

At Oilstainlab, building a car like the HF-11 isn’t just an engineering challenge — it’s a translation problem. The company’s 2,000-pound, powertrain-agnostic platform pushes unconventional design and manufacturing decisions that are difficult to visualize, explain and brand, especially for a small team working across a global supply chain. The challenge was to make those “maniacal” engineering choices legible and visceral, without the resources or timelines of a large automaker.

Nikita Bridan, co-founder and CEO of Oilstainlab, said the gap wasn’t creativity but scale. The team needed a way to connect radical engineering concepts to a coherent brand identity, while operating with tight deadlines and limited budgets. Traditional design and marketing workflows couldn’t keep pace with how quickly the project was evolving.

To close that gap, Oilstainlab began using AI as both a design and creative multiplier. On the visual side, the team developed a distinct, harsh flash-lighting aesthetic that mirrors the HF-11’s raw mechanical character. AI tools allowed that visual language to be applied consistently across images taken anywhere in the world—by professionals or amateurs—effectively “branding” the photography while dramatically expanding the company’s usable image library.

The result was a clearly defined visual identity that matched the car’s performance philosophy and lightweight engineering ethos. More importantly, it decentralized creative production. Anyone on the team could generate compelling, on-brand imagery without relying on a small set of specialists or costly reshoots.

For Bridan, the experience reinforced a practical lesson about AI’s role in creative work. AI wasn’t used to invent the car or fabricate imagery, but to amplify real design decisions and real photographs taken by real people. In that context, AI became less about replacing creativity and more about extending it—helping a small team move faster without losing control of its identity.

Have an AI use case to share with us? Email us at: ai.newsletter@newsweek.com

Context Window

■ The ABB FIA Formula E World Championship is partnering with Google Cloud to use AI—including Google Gemini and a GenAI-driven Strategy Agent—to enhance both operations and fan engagement, offering real-time tailored insights, predictions and explanations during live broadcasts while also optimizing logistics and back-office workflows. [Newsweek]

■ Meta is temporarily pausing teen access to its AI characters across apps as it works on updated versions with stronger safety controls, a move reflecting heightened industry focus on protecting minors from potential harms in AI interactions. [Associated Press]

■ Anthropic CEO Dario Amodei warns that AI is entering a volatile “adolescence,” arguing in a new essay that rapidly advancing systems are outpacing society’s ability to manage their risks and calling for stronger safeguards as autonomy increases.[Dario Amodei]

■ A coalition of U.S. state and federal judges launched the Judicial AI Consortium to study and prepare for both risks and opportunities as AI tools are increasingly used in courtrooms. [Reuters]

■ MIT researchers find that machine-learning models with the best average performance can nonetheless fail on large portions of new data, arguing that overly aggregated metrics hide critical reliability issues and suggesting more granular evaluation methods for trustworthy AI deployment. [MIT News]

■ At the 2026 World Economic Forum in Davos, executives and AI leaders said that while corporate AI investments remain high, many businesses are encountering friction in adoption and are under pressure to show measurable returns, with strategic deployment and workforce readiness emerging as key enterprise challenges. [Axios House]

Transfer Protocol

Vinoo Srinivas Murali, a seasoned sales and AI business leader, has been named chief revenue officer – data and AI at DigitalNet.ai, where he will lead global expansion and accelerate adoption of the company’s JanusAI platform and autonomous cybersecurity offerings.

Garrett Berntsen, a federal and industry technical leader, has been appointed chief artificial intelligence officer at Accenture Federal Services, where he will shape AI strategy and solutions for federal customers.

Hannah Poferl, former chief data officer at The New York Times, has been tapped by Universal Music Group to lead its data strategy as the company further integrates AI into its business.

Dan Koellhofer, a veteran product and SaaS leader, has been appointed chief product officer at Model N to accelerate AI-enabled product innovation as part of the company’s platform growth strategy.

Zachery Anderson, an analytics and data executive with prior leadership experience at NatWest Group, has joined JPMorgan Payments as chief data and analytics officer, where he will oversee data and analytics strategy for the company’s global payments business, including data platforms that support AI-driven capabilities.

Faraz Shafiq, an Amazon Web Services executive with more than 15 years of technology experience, has been appointed head of AI products and solutions at Wells Fargo, where he will oversee the vision, roadmap and development of enterprise-wide artificial intelligence-powered products.

Know someone on the move in AI? Send job change info to ai.newsletter@newsweek.com

Magic Moment

Dr. Marcus Weller, PhD | CEO, Deepinvent, Advisor, DARPA

...

“When we watched our AI system redesign its own architecture and then patent that breakthrough, it became clear that we had crossed a line. AI was no longer just helping humans innovate it was beginning to invent. That moment showed us that discovery is no longer limited by human bandwidth, and that the pace of progress is about to accelerate exponentially.”

Experience some AI magic? Tell us about it at ai.newsletter@newsweek.com

Read More on Newsweek.AI