Non classé
AI Risks Include Data Poisoning and Model Corruption
Published
1 an agoon
By
Ted Krantz, CEO of Interos
Interos, a company providing supply chain resilience and risk management software, emailed me to say that there was a supply chain risk everyone seemed to be ignoring – AI-related risks.
Companies use risk management software, like the Interos solution, to monitor and analyze supplier risk events in real time. These are big data platforms that monitor news sources and assorted databases from governments, financial institutions, ESG NGOs, and other sources to detect when an adverse event has occurred or may be about to occur.
It is well known that ChatGPT can hallucinate. Almost all supply chain software companies are talking about how they are incorporating generative AI into their solutions and how this could improve user interfaces. Most argue that when the UI is trained with the company’s own data, the risk of hallucination is small.
But what Interos is talking about is different. Not just the risk of generative AI hallucinations but the risks associated with all forms of AI. The AI-related risks include data poisoning and model corruption. These, Interos argues, pose significant challenges for organizations integrating AI into their operations. They are right, I don’t hear anyone else discussing this risk.
I interviewed Ted Krantz, the new CEO of Interos, to learn more. Mr. Krantz argues that with Cloud-based architecture, componentized software, and embedded analytics, there are significant flows of information across ERP and supply chain platforms. That information comes from inside the platform applications and increasingly from outside sources, like Interos. These platforms need high-fidelity signals that can be trusted.
The life cycle path of the data, Mr. Krantz continued, includes an input stage, the model, and the output. All three of those checkpoints have challenges and opportunities.
Garbage In, Garbage Out
Garbage in, garbage out refers to the data integrity problem on the input side. Solutions, particularly solutions that leverage public data like risk management applications, are very reliant upon the quality of the signal. Is the information from these websites factual? Is it fictional? “So, there’s a corruption component at the input level that all of us have to struggle through.” This is true whether it’s generative AI application or more traditional forms of artificial intelligence.
Some algorithms can help clean data. These can be simple logic that detects if it is a zip code field; there should be five digits. If the zip is only four numbers long, then it is wrong. Or data cleaning tools can suggest that both “P&G” and “Proctor & Gamble” probably refer to the same company.
However, numerous other input errors can occur, particularly surrounding something as complex as a global supply chain. But, Mr. Krantz continues, “This gets really complicated, really fast.” Interos provides risk scoring across six different types of risks. “Each one of those independent risk factors has individual, unique variables that can require manual intervention and scrubbing to adjust for and correct. There could be changes on the regulatory front. For example, over 15,000 companies were added to the US restricted entities list in 2023 and 2024.
Or perhaps a piece of legislation’s go-live date is delayed, which changes the scoring. “It’s a hornet’s nest of literally countless potential corruptions at the input level that you’re constantly adjusting to. So, the primary point here is that for the foreseeable future, we need a team that is constantly checking the data.” The team, the CEO explains, is constantly “banging” on the system to try to detect errors as well as interacting with customers who think some of the scores may be wrong. “This is endless. It’s a beast. People don’t just get replaced here. They actually get put in more strategic positions.”
AI Model Corruption
The AI models can also become corrupt. “The complexity at the model level is the orchestration of the private signals, the company signals, what signals are superseding by others, and getting that calibration correct.” For Interos, the AI model calculates the Interos risk score on a 0 to 100 scale. There are green, yellow, and red indicators, and maps and monitoring capabilities attached to the scores.
At the model level, one set of issues surrounds how that score is framed. “What are the variable weightings associated with that score?” How much weight should be given to each variable that makes up the score? “Much like at the input level, we need a team around this that constantly calibrates how the score should be calculated.” And like at the input level, ongoing collaboration with clients is necessary to ensure that the scoring mechanism is accurate. There is always a “human in the loop. If anyone’s saying that they don’t have that, they’re just not being truthful.” For example, online news items can generate event data. But to create real-time maps surrounding that risk often requires humans to tune the algorithm.
Whatever score is generated, “customers are naturally going to challenge that score.” We have to have a way for us to frame the integrity of the I score that is strictly an unbiased industry purview based on the data that we see.” For example, a customer might see a cyber risk score of 70. Sophisticated customers seek to understand how that score was generated and argue that if the score were calibrated differently, their score would be higher. And that customer might be right. There has to be an element of collaboration around the risk model.
One thing Interos is working toward is giving the customer the ability to weigh the parameters themselves. For example, a supplier’s risk score might be based partly on the FICO creditworthiness score generated by the Fair Issac company. In the future, Interos customers may decide to give that variable either more or less weight.
Interos’s CEO points out that the risk models have different levels of complexity. Some of the data, like FICO scores, is “quasi-historical.” In some cases, like predicting storm paths and which suppliers might be impacted, it is a real-time prediction. For more complex models, Interos needs to move more slowly before allowing customers to change the parameters.
AI Outputs Can Be Corrupt
Finally, AI outputs can be corrupted. One risk here is the potential loss of intellectual property. This is the idea that a hacker or malign government might be able to find an entry point to a company’s application and view, corrupt, or lock up the data. All enterprise software suppliers need robust cybersecurity in place.
Interos just published a report called 5 Supply Chain Predictions You Need to Know in 2025. Interos predicts that traditional cyber attacks – malware, ransomware, phishing, etc. – will continue in 2025, but they warn that we need to be on the lookout for more disruptions to the physical infrastructure that is foundational to our digital world. Geopolitics rivalries underpin the potential for significant cyber disruptions to the hardware and software we depend on to make our world go round. Increasingly, enterprise applications run on Public Clouds. An attack that takes down a public cloud platform does not just affect one company; it affects numerous companies.
The post AI Risks Include Data Poisoning and Model Corruption appeared first on Logistics Viewpoints.
You may like
Ocean freight forwarding is an $80+ billion market bogged down by the manual processes related to booking management, documentation services, and the coordination labor that holds it all together.
When working with a freight forwarder, you’re buying three things bundled together:
Carrier relationships — access to capacity, negotiated rates, allocation commitments.
Operational data — knowing which carrier fits a given lane, what documents a particular trade corridor requires, how to handle an exception when a booking gets rejected.
Coordination labor — the booking itself, the documents per container (industry estimates range from 9 to 18 depending on the corridor), the re-keying of data across disconnected systems, the email chains chasing confirmations and clearances.
Shippers have always paid for the bundle because you couldn’t get one piece without the others, but that’s changing.
Where the bundle comes apart
Travel agents used to bundle airline relationships, destination expertise, and the labor of putting trips together into a single fee. Aggregator platforms unbundled the pieces, and the booking layer went first because that’s where the volume was. Ocean freight forwarding is in the same position. More than digitizing booking, though, AI is automating it.
The bulk of the volume and labor cost for freight forwarders is tied up in rate comparisons across dozens of carriers, document preparation and routing by trade lane and commodity classification, booking execution against pre-negotiated contracts, and exception triage on rejected bookings.
But this is all high-volume, rule-governed, multi-system coordination where speed and consistency matter more than creativity. Exactly the type of work that AI agents are well-equipped to handle.
Platforms can now ingest a rate agreement, parse surcharges and FAK provisions into a digital rate profile, compare carriers on cost, transit time, and schedule reliability, and execute a booking based on pre-defined parameters, without a human in the loop.
Automating the entire order lifecycle
Every dollar of margin exposure in ocean freight traces back to a decision made without complete information. That means that every action must be rooted in live network data across shipment flows, carrier performance, and insight from inventory and order systems. A platform with that intelligence can automate and accelerate the full workflow from detecting a supply shortfall, selecting a carrier, booking the container, managing the documents, tracking the shipment, and handling exceptions.
A shipper stitching together a rate tool from one vendor, a booking portal from another, a document system from a third, and a visibility feed from a fourth gets digitization. They get a slightly faster version of the same manual process. The full picture still lives in a person’s head, and the handoffs between systems still require human coordination.
While freight forwarders and other intermediaries are also investing in AI, they’re primarily automating their own coordination labor before someone else absorbs it. But they can’t replicate the data advantage of a platform that sits across the entire supply chain.
A forwarder automating its booking desk draws on its own transaction history. A point solution built specifically for ocean booking draws on booking data. A platform processing millions of supply chain events daily across orders, inventory, carrier performance, and live shipment status, has a different signal base entirely. Carrier selection informed by real-time schedule reliability, live network disruption, and your actual inventory positions is structurally more accurate than carrier selection informed by historical rate tables.
The shrinking intermediary layer
The moats around freight forwarders’ profit margins are eroding, and the lines between legacy endpoint solutions are blurring. High-complexity corridors and specialized commodities still need human expertise, but the bread-and-butter containerized freight that makes up the bulk of forwarder revenue is the volume where automated workflows shine.
Meanwhile, software providers will have a hard time selling dashboards and chatbots to specific teams compared to AI-native platforms offering a single operating system across all supply chain operations, and serving downstream stakeholders.
The question for forwarders is how long they can keep patching automation onto a fragmented architecture with a booking tool here, a document system there, people bridging the handoffs in between. And how much revenue sits in structured, repeatable work that a connected platform absorbs?
For shippers, the choice is whether to invest in a platform that automates the order-to-delivery and exception lifecycle, or keep paying others to hold the pieces together. The second option is a decision to fund the intermediary layer sitting between them and their own data.
The post The Freight Forwarder Moat Is Getting Shallower appeared first on Logistics Viewpoints.
Non classé
Supply Chain and Logistics News Week of May 7th 2026
Published
19 heures agoon
8 mai 2026By
The logistics and supply chain landscape is undergoing a fundamental transformation as industries move from rigid, low-cost models toward strategies defined by agility and resilience. This week’s roundup explores how major players are navigating this shift, from Amazon’s bold move to offer its massive infrastructure as a standalone service to Ford’s strategic manufacturing reset in the EV sector. We also dive into the critical human element in modern cost engineering, the logistical reimagining of energy corridors due to geopolitical risks, and the new AI-driven tools closing the gap between inventory detection and real-time execution. Together, these developments highlight a common theme: the pursuit of flexibility and data-driven intelligence in an increasingly unpredictable global market.
Top Supply Chain Stories from this Week:
Modern Cost Engineering Evolution: Rewiring the Human Element for Supply Chain Resilience
In the latest shift for cost engineering, the focus is moving beyond purely digital tools to address the critical human element required for true supply chain resilience. As industrial organizations transition from traditional backward-looking estimates to modern “should-cost” methods powered by AI and digital twins, the real challenge lies in workforce transformation. Success in this new landscape requires a significant cultural shift, moving away from isolated departmental silos toward cross-functional collaboration. By reskilling traditional estimators to act as strategic consultants—capable of interpreting material science and operational constraints—companies can evolve from simple price negotiation to collaborative manufacturing improvements that ensure mutual profitability and long-term stability.
Hormuz Risk Is Redrawing the Supply Chain Geography of Energy
Geopolitical instability in the Strait of Hormuz is forcing a fundamental shift in energy logistics, moving the industry away from lowest-cost network design toward a risk-adjusted model. With the waterway handling roughly 20% of the world’s oil and liquefied natural gas, repeated disruptions have transformed infrastructure like pipelines, storage terminals, and deep-water ports outside the Persian Gulf into high-value strategic assets. Nations and corporations are no longer viewing these as simple logistics nodes, but as essential escape routes that provide the optionality and recovery time needed to withstand chokepoint failures. This selective redesign of the global energy map signals a new era where geography and physical redundancy are the primary drivers of supply chain resilience.
Ford’s Manufacturing Reset Shows How Automakers Are Rebuilding the EV Supply Chain
Ford’s manufacturing pivot represents a fundamental shift from aggressive electric vehicle expansion toward capital discipline and supply chain flexibility. By taking a $19.5 billion write-down and restructuring battery joint ventures, the company is moving away from rigid, single-purpose production lines in favor of multi-energy platforms that can adapt to fluctuating demand for hybrids and EVs. A key component of this reset is the repurposing of battery manufacturing assets in Kentucky and Michigan for stationary energy storage and data center support. This strategy transforms these facilities into flexible energy infrastructure rather than just automotive supply nodes. Ultimately, Ford is signaling that the next phase of the market will be defined by the ability to manage uncertainty through cross-functional asset utilization and a focus on manufacturing-driven affordability.
How FourKites Connects Stockout Detection to Freight Execution in Minutes
FourKites has launched a unified solution that bridges the gap between stockout detection and freight execution, reducing resolution time from hours to less than five minutes. By integrating its Inventory Twin and Booking Connect AI, the platform eliminates the traditional “manual scavenger hunt” where planners had to jump between ERPs and carrier portals to resolve inventory gaps. The system uses decision intelligence to identify stockout risks up to six weeks in advance and provides ranked recommendations for corrective transfers based on cost, speed, and carrier performance. This closed-loop workflow allows planners to execute optimized shipping options with a single click, addressing the massive financial impact of inventory distortion and reducing the need for expensive, unplanned expedited shipping.
Amazon Launches “Supply Chain Services” Leveraging its Global Logistics Network
Amazon has officially launched Amazon Supply Chain Services (ASCS), a move that decouples its massive logistics infrastructure from its retail marketplace to serve as a standalone utility for all businesses. Similar to the trajectory of Amazon Web Services (AWS), the platform opens up Amazon’s multimodal freight, automated warehousing, and last-mile parcel delivery networks to companies regardless of whether they sell on Amazon. Major early adopters like Procter & Gamble, 3M, and Lands’ End are already leveraging the service to move everything from raw materials to finished products. By consolidating fragmented logistics contracts into a single automated interface, Amazon aims to use its scale—currently moving 13 billion items annually—to provide businesses with end-to-end visibility and 96.4% on-time delivery rates, signaling a significant new challenge to traditional 3PLs and carriers like FedEx and UPS.
Song of the week:
The post Supply Chain and Logistics News Week of May 7th 2026 appeared first on Logistics Viewpoints.
Non classé
How FourKites Connects Stockout Detection to Freight Execution in Minutes
Published
1 jour agoon
7 mai 2026By
FourKites is bridging the gap between identifying a problem and solving it. With the integration of Inventory Twin and Booking Connect AI. Traditionally, supply chain planners have been stuck in a manual scavenger hunt whenever a stockout alert surfaced, jumping between ERPs to find surplus stock and carrier portals to secure freight. This fragmented process typically took hours, often forcing companies to rely on expensive, last-minute expedited shipping or facing steep On-Time In-Full (OTIF) penalties to avoid customer dissatisfaction. By unifying these disparate data streams, the new solution allows teams to detect risks two to six weeks in advance and execute corrective transfers from a single, seamless workflow.
The impact on operational efficiency is significant, reducing the resolution time from detection to execution from several hours to less than five minutes. Instead of just receiving a warning, planners are presented with recommendations powered by Decision Intelligence that include the fastest, cheapest, and most optimal shipping options based on real-time carrier performance data. This closed-loop system directly addresses the 1.73 trillion dollar global issue of inventory distortion and aims to eliminate the 15-25 hours planners previously spent on manual coordination.
By keeping a human in the loop to select the best recommendation with a single click, FourKites ensures that exceptions are resolved without ever leaving the platform. This integration helps protect freight budgets, where unplanned expedited shipping often consumes up to 48% of total spend. This launch represents a shift from reactive firefighting to proactive execution, allowing teams to move away from costly safety stock and focus on high-value responsibilities. Supply chain planner responsibilities are changing with the continued developments of AI and the de-siloing of disparate systems.
FourKites is a supply chain technology provider that operates a global real-time visibility network tracking over 3.2 million shipments daily across 200 countries and territories. By integrating data from 1.1 million carriers across all modes (road, rail, ocean, and air), the platform uses AI-powered “digital workers” to automate exception resolution and provide predictive insights. More than 1,600 global brands, including leaders in the CPG and Food & Beverage sectors, trust FourKites to transform their logistics from reactive tracking into proactive, intelligent orchestration.
Read the full ARC brief breaking down the new FourKites solution here: https://www.fourkites.com/research/arc-advisory-stockout-detection-freight-execution/
The post How FourKites Connects Stockout Detection to Freight Execution in Minutes appeared first on Logistics Viewpoints.
The Freight Forwarder Moat Is Getting Shallower
Supply Chain and Logistics News Week of May 7th 2026
How FourKites Connects Stockout Detection to Freight Execution in Minutes
Walmart and the New Supply Chain Reality: AI, Automation, and Resilience
Ex-Asia ocean rates climb on GRIs, despite slowing demand – October 22, 2025 Update
13 Books Logistics And Supply Chain Experts Need To Read
Trending
-
Non classé1 an agoWalmart and the New Supply Chain Reality: AI, Automation, and Resilience
- Non classé7 mois ago
Ex-Asia ocean rates climb on GRIs, despite slowing demand – October 22, 2025 Update
- Non classé9 mois ago
13 Books Logistics And Supply Chain Experts Need To Read
- Non classé3 mois ago
Container Shipping Overcapacity & Rate Outlook 2026
- Non classé3 mois ago
Ocean rates ease as LNY begins; US port call fees again? – February 17, 2026 Update
- Non classé6 mois ago
Ocean rates climb – for now – on GRIs despite demand slump; Red Sea return coming soon? – November 11, 2025 Update
-
Non classé1 an agoAmazon and the Shift to AI-Driven Supply Chain Planning
- Non classé2 ans ago
Unlocking Digital Efficiency in Logistics – Data Standards and Integration
