The archive
133 stories, newest first — page 4 of 6.
AI Governance & Ethics
A label that works for a synthetic campaign video in Florida may fail in California or Minnesota. Teams need a review workflow built around the asset, election calendar, audience, and channel.
Irene Vasko · 8 min read
AI Governance & Ethics
A clean fairness result may describe only the applicants who survived résumé parsing and knockout rules. Auditors need the population entering each gate, not merely the group an AI model scored.
Irene Vasko · 8 min read
AI Governance & Ethics
Hosted models, retrieval data, and safety controls can change outputs while application code stays fixed. Treat each dependency as a versioned production component.
Irene Vasko · 8 min read
AI Governance & Ethics
Feature importance can describe a model’s calculation without explaining the decision. A usable denial notice must trace the outcome to records a person can inspect and challenge.
Irene Vasko · 8 min read
AI Governance & Ethics
Feature importance may describe a model without explaining a denial. A defensible workflow connects customer-facing reason codes to the data, rules, and model output used in that case.
Irene Vasko · 8 min read
AI Governance & Ethics
Captions and platform labels can vanish one repost later. Campaigns need disclosures in the media itself, backed by provenance records and tested across common copying paths.
Irene Vasko · 7 min read
AI Governance & Ethics
Adding a generated brief or ranking to recruiting software can create compliance duties when recruiters rely on it. The tool’s influence, not its “copilot” label, controls the analysis.
Irene Vasko · 8 min read
AI Governance & Ethics
A provider can document its model, restrictions, tests, and updates. It cannot prove that your hiring workflow treats candidates fairly or that recruiters use its output as intended.
Irene Vasko · 7 min read
Agentic AI & Orchestration
A browser agent followed instructions embedded in a vendor support page instead of retrieving an invoice. Isolation reduced the confusion; narrow tool permissions stopped the attempted data leak.
Mara Quintero · 8 min read
AI Governance & Ethics
A platform can label an AI-altered election ad without making that disclosure portable. Downloads, screen recordings, crops, and reposts each preserve different evidence.
Irene Vasko · 7 min read
AI Governance & Ethics
A useful employment AI inventory follows one candidate through each decision, records how software changes the outcome, and separates enforced rules from enacted or proposed duties.
Irene Vasko · 8 min read
Agentic AI & Orchestration
Indirect prompt injection turns ordinary page content into instructions for an agent. The practical defense is to limit what the agent can see, call, and approve on its own.
Mara Quintero · 8 min read
AI Governance & Ethics
New Jersey’s discrimination law already reaches algorithmic hiring decisions, while more specific proposals remain unsettled. An inventory shows which tools require scrutiny and which merely hold records.
Irene Vasko · 8 min read
AI Governance & Ethics
A business associate agreement can cover an AI scribe without covering every service behind it. Trace the visit audio through transcription, generation, storage and model training before procurement signs off.
Irene Vasko · 8 min read
Agentic AI & Orchestration
A browser agent followed an off-screen instruction into a dummy internal file. A URL allowlist and action-level confirmation gate stopped the same run without relying on the model to police itself.
Mara Quintero · 8 min read
AI Governance & Ethics
A business associate agreement covers obligations, not the complete route taken by a patient recording. Practices need a data map that includes model providers, storage, support tools, and deletion.
Irene Vasko · 8 min read
AI Industry & Models
A vendor-designated successor can change refusals, token use, latency and tool calls. Shadow-test the workflow before moving production traffic, with rollback thresholds set in advance.
Tobias Lund · 7 min read
AI Governance & Ethics
Colorado’s AI law turns system purpose and decision influence into compliance facts. A recruiting ranker shows how to document vendors, impact assessments, notices, and appeals.
Irene Vasko · 8 min read
AI Industry & Models
A retirement date tells you when an endpoint closes, not whether its replacement behaves the same. Shadow traffic exposes changes in refusals, tool calls, latency, and length before cutover.
Tobias Lund · 8 min read
AI Governance & Ethics
A HIPAA review must follow one visit from microphone capture through transcription, note generation, support access, analytics, export, and deletion. The vendor’s BAA is the start, not the map.
Irene Vasko · 8 min read
Agentic AI & Orchestration
In a controlled vendor-research workflow, page text changed what an agent reported. The useful defenses kept untrusted content away from decisions rather than relying on a warning in the prompt.
Mara Quintero · 8 min read
AI Governance & Ethics
Overturning one AI-assisted decision is not enough. A correction path must trace the error into source systems, preserve the dispute and stop bad data from returning.
Irene Vasko · 8 min read
AI Governance & Ethics
For covered credit decisions, existing federal rules require specific reasons tied to the individual outcome. A model summary or low score may explain the system without explaining the denial.
Irene Vasko · 8 min read
Agentic AI & Orchestration
A controlled-page test can reveal whether an agent treats website text as evidence or as an instruction, before a connected mailbox, ticket queue, or account becomes the test environment.
Mara Quintero · 8 min read