Articles and Activities / All sites
Articles and
activities.
Everything I publish on contestability.ai and dghumana.com, and the work of Algo BirdDog Analytics at algobirddog.com, in one list, newest first: short articles, commentary, consultation and rulemaking responses, research examples, and research activities.
Each entry is marked with its site and links to it there. The list updates itself as new work appears on each site.
Short article
A summary is already an interpretation. What Spain’s Constitutional Court will not hand to a machine
What was said On 12 November 2026 the Spanish outlet Moncloa reported remarks by Juan Carlos Campo, a magistrate of the Constitutional Court who is leading its digital transformation. The court is studying how…

Short article
When explaining the AI lowers the wish to appeal, has anyone been given grounds?
What was reported On 10 November 2026 University College Dublin’s Smurfit School and Trinity College Dublin issued a press release about a study by Christopher Greene, Brian Barry and Marius Claudy, published online on…

Short article
Europe’s regulators will test whether a hiring rejection can be explained. What would the rejected candidate need to be able to find out?
What was announced On 11 November 2026 the European Data Protection Board EDPB said, at the close of its latest plenary, that its sixth coordinated enforcement action will concern automated decision-making in recruitment, including…
Short article
A signature is not a review. What the new Ibero-American AI standards count as human intervention
What was approved On 12 November 2026 the Spanish data protection agency announced that the Ibero-American Data Protection Network, whose permanent secretariat it holds, had unanimously approved standards on the use of personal data…
Short article
The Digital Fairness Act proposal lets a court demand a company’s evidence. Who has to notice first?
What was proposed On 11 November 2026 the European Commission adopted a proposal for a Digital Fairness Act, COM 2026 1100. It would replace the Unfair Commercial Practices Directive with a regulation and add…
Short article
Tell the feed what you want, and your words may count as one more signal. What is a person’s own word worth?
What was said On 13 October 2026 Meta introduced “Your algorithm” for Facebook. Users can add topics for Feed and Reels, and after tapping “Interested” or “Not interested” they may be asked to say,…
Short article
Vietnam’s bank-AI circular promises a human review of complaints. What will the reviewer have in hand?
What takes effect On Sunday 15 November 2026, Circular 53/2026/TT-NHNN of the State Bank of Vietnam takes effect. It was issued on 30 September and sets safety, risk and deployment conditions for AI systems…
Short article
A decision with no sentence in it. What is lost when a machine answers only in probabilities?
What was announced On 11 November 2026 Google DeepMind introduced the Gemini Decisions API, powered by a model called Gemini Decision 1. We read Google’s announcement and its documentation. The model does not write…
Discussion paper response submitted
Making agentic banking decisions contestable
A response to the Hong Kong Monetary Authority’s Discussion Paper on Agentic A.I. Governance in the Banking Sector proposes a practical test for customer contestability. A hypothetical loan application in two versions separates a classification error that ordinary records resolve…

Research example
A corrected income figure is not yet a reconsidered credit decision
A fictional French consumer-credit example. A competent human reviewer resolves an income-conversion objection without giving the borrower a calculation extract, and the lending decision remains open. Includes the six-page packet and a French version.

Workshop response submitted
Making AI-assisted insurance claims open to meaningful challenge
Written comments to the California Department of Insurance use a fictional roof claim to examine evidence, causal inference, human review, and vendor accountability. Six proposed safeguards distinguish existing legal duties from NIC research and implementation proposals.

Research example
Correct calculation, contested rule: a California employment example
An employment example, three employee notices, and internal review records distinguish correct calculation from a justified disciplinary decision. Includes the 11-page packet and a legal scope companion.

Research activity · Decision pending
Penn–Georgetown workshop abstract submitted
Kerry Thornhill submitted an abstract on the contestability of automated decisions to the 5th Annual Penn–Georgetown Digital Ethics Workshop. The submission forms part of the project’s ongoing research into the conditions for meaningful challenge and institutional accountability. A decision is…
UK consultation response
Making agentic AI decisions open to meaningful challenge
A response to the call for evidence on agentic AI from the Information Commissioner's Office, the UK's data protection regulator. Submitted November 13, 2026.

Short article
Whose Reasons Count?
On the Stanford Encyclopedia of Philosophy's October 2026 revision of "Feminist Social Epistemology."

Rulemaking comment
Making review rights usable: Colorado ADMT rulemaking
A filed comment proposing decision-specific reasons, usable reconsideration, and records of what human reviewers actually considered. Includes the complete submission. Article added November 15, 2026.

Commentary
Who Decides What Counts as a Reason?
Habermas argues that rational agreement depends on the force of reasons. When people seek understanding through argument, they treat claims as open to criticism and their acceptance as something that requires justification. An institution’s…

Commentary
Claude Questioned Reality. But Who Taught It to Ask?
Anthropic’s AI solved a decades-old cryptographic mystery, then questioned whether its discovery was real. The achievement is impressive. But interpreting the model’s apparent skepticism requires a closer look at the conditions under which it…

Research preprint
From Notice to Contestable Grounds
Paper A develops the Necessary Informational Conditions framework: what an affected person must be able to investigate when meaningful contestability is required or promised. Read the paper, its abstract, and publication details.
