An independent publication about how artificial intelligence is changing technology, business, work, science, security, law, culture, and everyday life.
Evidence before certainty.
On this page: What we do · Editorial standard · Updates · How we use AI · Who is behind it · Independence · Methodology · Contact and corrections
We look beyond product launches and headlines to examine what AI is actually doing, what the available evidence supports, where important claims remain uncertain, and what those developments mean in practice.
Our coverage includes AI agents, models, infrastructure, research, cybersecurity, privacy, regulation, healthcare, finance, education, work, creativity, entertainment, geopolitics, and the wider social impact of artificial intelligence.
What AInvasion does
AI is developing faster than most people can reasonably track. New models launch. Companies announce deployments. Governments introduce regulations. Research produces new findings. Products promise major productivity gains.
But an announcement is not the same thing as evidence. AInvasion exists to examine the gap between the two.
What we publish
- Independent AI analysis grounded in available evidence.
- Research explainers that put technical and scientific developments into context.
- Industry analysis covering the practical effects of AI on organizations and workers.
- Coverage of AI security, privacy, law, regulation, safety, and accountability.
- Evaluations and comparisons of AI tools based on documentation, pricing checks and calculations, with the limits of each check stated (see How We Test).
- Practical guides, calculators, datasets, and other resources that help readers use AI information rather than simply consume it.
We do not assume that every AI development is revolutionary. Sometimes the evidence shows that a technology works. Sometimes it shows that a claim is narrower than the headline suggests. Sometimes the evidence is simply not strong enough to know yet. All three are legitimate conclusions.
Our editorial standard
Evidence before certainty
Important factual claims should be traceable to an underlying source. Depending on the subject, we may use:
- Peer-reviewed research
- Government agencies and regulators
- Official company documentation
- Technical documentation
- Financial filings
- First-party datasets
- Independent testing by third parties
- Established research organizations
- Industry reports
- Reliable secondary reporting for additional context
Primary and authoritative sources are preferred whenever they are available. We also distinguish between a source’s claim and what the evidence independently establishes.
Facts, analysis, and uncertainty are different
AInvasion separates three things:
- What is documented. What a source, dataset, company, regulator, research paper, or test actually establishes.
- What the evidence suggests. An interpretation supported by available evidence but still subject to uncertainty.
- What we think it means. Editorial analysis of the implications, limitations, and practical consequences.
When evidence is incomplete, contradictory, geographically limited, commercially biased, or too early to generalize, we say so. Our longer analysis pieces may also use these confidence labels: Established, Probable, Speculative.
The complete process is documented on our Methodology page.
We update when the evidence changes
Products change. Pricing changes. Regulations change. Models improve. Companies discontinue projects. New research can overturn earlier assumptions.
Publication is therefore not always the end of the editorial process. When a material fact changes, we update the relevant article and identify significant corrections or changes where appropriate. An article should remain useful after publication, not simply remain unchanged.
How we use AI
AInvasion uses Claude, an AI assistant made by Anthropic, as part of its editorial workflow. Claude does much of the research discovery, source-checking, recalculation, drafting, editing, data analysis, and technical work. The editor sets the standards, reviews the output, and decides what is published.
AI assistance does not replace editorial responsibility. Important claims and statistics are checked against their underlying sources. AI-generated wording is revised or removed when it goes beyond what the available evidence supports.
What this means for verification, including what “tested” does and does not mean on this site, is explained on How We Test.
Tom Morgan is the Founder & Editor of AInvasion and is responsible for the publication’s editorial direction, sourcing standards, and corrections. AInvasion examines what AI systems actually do, what the available evidence supports, and where important claims remain uncertain. Product assessments distinguish published specifications, third-party testing, and our own documented checks, and uncertain conclusions are identified as such.
Who is behind AInvasion?
Tom Morgan, Founder & Editor. His editorial focus includes AI adoption, technology strategy, AI safety, and the gap between how emerging technologies are presented and how they perform in real-world settings.
AInvasion does not present Tom Morgan as an authority on every specialized field covered by the publication. Where a subject requires specialist expertise, the relevant research, institutions, technical documentation, or subject-matter sources are identified rather than substituted with unsupported claims of personal expertise.
The publication’s credibility is therefore based on a combination of named editorial accountability, transparent sourcing, structured verification, and a clear correction process.
Independent by design
AInvasion is editorially independent. Companies and products may be examined positively, negatively, or inconclusively depending on the available evidence. Where relevant, we distinguish between:
- company claims;
- independently verified information;
- third-party research;
- our own checks (recalculation and code runs, see How We Test);
- editorial interpretation.
A company cannot purchase a favorable editorial conclusion. Affiliate relationships, sponsorships, or other commercial relationships do not determine the conclusions of our analysis.
What we don’t assume
We do not assume that:
- every benchmark reflects real-world performance;
- every AI pilot produces meaningful ROI;
- every new model represents a major breakthrough;
- AI will automatically replace entire professions;
- AI will automatically solve complex social or economic problems;
- skepticism and enthusiasm are mutually exclusive.
We also do not treat uncertainty as a weakness. Sometimes the most accurate conclusion is: we do not know yet.
Why AInvasion exists
AI is becoming part of software, infrastructure, workplaces, healthcare, financial services, education, entertainment, shopping, security, and everyday life. There is more information about AI than ever before. There is not necessarily more clarity.
AInvasion exists to help close that gap. We research the claims, examine the evidence, explain the implications, and identify what remains uncertain.
Research, methodology & corrections
For a detailed explanation of how AInvasion verifies claims, evaluates sources, labels uncertainty, handles corrections, and uses AI in its editorial workflow, see our Methodology page. For what each verification label means and what we do not test, see How We Test.
Contact and Corrections
If you find an inaccurate claim, outdated statistic, missing source, or misleading statement, please contact AInvasion with the relevant article URL and supporting evidence where possible.
We review substantive corrections against the available evidence. When a material error is confirmed, we correct the article and explain significant changes where appropriate.
For corrections, research questions, source submissions, or other editorial matters, visit our Contact page or write to support@ainvasion.com.
AInvasion follows one principle: evidence before certainty. When the evidence changes, our conclusions should change with it.
Tom Morgan Founder & Editor, AInvasion
Shorter notes on X: @rkmlm1976. Code and side projects on GitHub: rkmlm. New articles land in the newsletter below, with no spam and no sponsored placements.