Who he is
Rahim Hirji is the author of SuperSkills: The Seven Human Skills for the Age of AI, published by Kogan Page in 2026, and the founder of The SuperSkills Intelligence Company. He is based in London and speaks worldwide, in English.
He spent twenty years building education technology before he wrote about it. He founded EtonX, the online learning business later acquired by Eton College, and led Quizlet’s international expansion across sixty countries, working out of a company headquartered in San Francisco. The argument rests on research across more than 200 organisations in 30 countries since 2019, published in full at the research.
The argument
AI rarely takes a whole job. It takes the first draft, the initial analysis, the routine review. Those were also the tasks through which people built the judgement that made them senior. The work still ships, the output often improves, and nothing triggers an alarm.
The question for a leadership team is therefore about mode rather than adoption. Is the organisation drifting into AI through a thousand small reasonable decisions nobody quite made, or designing it by deciding in advance where human judgement has to remain? Drift asks nothing of you. That is what makes it the default.
What North America has instead of a law
Every European audience assumes the United States is the unregulated one. The assumption is half right and the other half is more interesting. There is no federal AI statute in either country. What there is instead is a patchwork, and three pieces of it speak directly to whether a human being is still deciding anything.
New York City passed the first AI hiring law in the world. Local Law 144 has required bias audits of automated employment decision tools since 1 January 2023, and its rules define the trigger as a system used, among other things, to use a simplified output to overrule conclusions derived from other factors including human decision-making. On 2 December 2025 the New York State Comptroller published audit 2024-N-6 on how that law is being enforced. The city’s department surveyed 32 companies and found one issue. The auditors reviewed the same 32 and found at least seventeen.
Illinois amended its Human Rights Act to cover AI in employment decisions, in force since 1 January 2026, and instructed its Department of Human Rights to make the notice rules. Eight months on, the rules do not exist. The duty is live and the instruction manual is not.
The federal executive branch runs on a memorandum rather than a statute, and it changed hands. OMB M-25-21, issued 3 April 2025, rescinds and replaces M-24-10. It requires human oversight, intervention and accountability for high-impact uses, and a route to timely human review and appeal. Read the whole document and the words automation bias and over-reliance do not appear in it once. It mandates the safeguard without naming the way the safeguard fails, which is the subject of human in the loop is not a safeguard.
Canada has no AI law at all, and one province wrote the strongest right on the continent
Canada’s federal AI bill died. The Artificial Intelligence and Data Act fell with Bill C-27 on prorogation on 6 January 2025 and has no successor. Ontario requires an employer to state in a job posting that AI was used to screen applicants, and the ministry’s own guidance says it is enough to state that it is used.
Quebec did something else. Section 12.1 of its private-sector privacy act, in force since 22 September 2023 and unamended, gives a person subject to a fully automated decision the right to submit observations to a member of the personnel of the enterprise who is in a position to review the decision. Not a right to an explanation. A right to argue with a named human who can change the answer. It is set out at the Montreal page.
The number a North American board should have
On 1 September 2026 the Federal Reserve Bank of New York published its August survey supplements covering firms in New York State and northern New Jersey. AI use among service firms had reached 61 per cent, from 40 per cent a year earlier and 25 per cent the year before that. Manufacturers reached 51 per cent.
The employment findings are the part worth carrying into a board meeting, because they cut against the headline everyone expects. Four per cent of service firms had laid anybody off because of AI. Fifteen per cent had hired fewer people than they otherwise would, and thirteen per cent had hired more. And the median share of a firm’s own workers actually using the tools was 17 per cent in services and seven per cent in manufacturing. Adoption is broad and shallow at the same time.
The Bank then reported, in its own words, that companies emphasised training employees to verify AI outputs, understand potential biases, follow data security protocols, and avoid over-reliance on the technology. A regional central bank, surveying its own district, naming the failure mode. That is the keynote, said by somebody else.
One caution to carry with it: no United States or Canadian government body measures what AI is doing to entry-level hiring. Both the New York Fed and Statistics Canada point outward to academic work for that. The evidence on the missing rungs is at missing rungs, graded for what it does and does not prove.
Where and how he speaks in North America
For boards, executive committees, corporate and association conferences, HR and CHRO conferences and leadership offsites. Delivery is in English. Rahim travels from London, and where travel is not practical the same session is delivered online. He can travel to the region or deliver virtually across all time zones.
Three keynotes, all versions of one argument. Drift versus Design is the signature talk for boards and executive teams. We Are Superheroes covers the seven human skills that grow more valuable as the tools spread. WTH, What the Human is a judgement test for CHROs and leadership teams, run live on the day, December to March only. Full descriptions are at keynotes.
He is the wrong choice for a tools demonstration, a vendor pitch or a technical AI briefing, and for any brief that needs an optimistic message with the difficult part removed. If that is the brief, the guide to choosing an AI keynote speaker names four other kinds of speaker and when to pick each.
Four jurisdictions, one recurring worry
From Local Law 144 to section 2-102(L) and Quebec's review rights, North American instruments circle the same fear: a human who signs without deciding. I was GM International at Quizlet, a US company, and I have advised companies across the US, so these rules are familiar ground. A group with exposure in several jurisdictions gets one map of decisions that holds in all of them, with a named owner for each.
Rahim was GM International at Quizlet and has advised companies across the US. He has written Box of Amazing, a free weekly letter, since 2017. Keynotes · Advisory · Box of Amazing
The three keynotes, and which one fits
Keynotes run 30 to 90 minutes, in person or virtual. A 30-minute main-stage version of each talk is available for conferences, and the full talk runs 45 to 60 minutes. The sector examples are chosen for each audience, around the same argument. All three, with the showreel, are at keynotes, and the programme copy is at the speaker pack.
Drift versus Design is built from your room. Your audience answers five anonymous questions beforehand, their answers shape the talk, and afterwards you get a one-page readout with the three decisions to look at first.
Also speaking nearby: Chicago · Montreal · Washington DC · New York.
Drift versus Design
Drift versus Design: why most organisations hand their judgement to AI without deciding to, and how to design your way through instead.
From Local Law 144 to section 2-102(L), state and provincial rules keep naming the same risk. Executive teams across North America turn that patchwork into one design question.
We Are Superheroes
We Are Superheroes: AI is the suit. The human decides.
Change Readiness opens the North American version, for teams whose tools change every quarter, with the seven skills set out so a mixed conference can spot them in an ordinary week.
WTH (What the Human)
WTH (What the Human): a live test of a leadership team's own judgement.
Rules in New York, Washington, Illinois and Quebec differ in letter and agree in worry. Directors with exposure across them bring a case from their business and check it against that shared concern.
City pages in this region
- New York The first AI hiring law in the world, and the state audit that found it unenforced.
- Montreal The right to argue with a named human who can overturn the machine.
- Washington DC Federal guidance named automation bias in March 2024 and stopped naming it thirteen months later.
- Chicago A civil rights duty in force since January 2026, and the rules the statute ordered still do not exist.
