Human-Agency Technology Test
Judge automation by whether it expands agency, meaning, and shared benefit
- Difficulty
- Moderate
- Time to result
- ~ongoing to results
- Steps
- 6
- Confidence
- 96%
In response to Bartlett's concern about rapid AI displacement, Poilievre says he has principles rather than a detailed forecast. His test asks whether technology gives people more agency, freedom, control, and ability to pursue meaning, or instead replaces their utility and leaves them without ways to contribute. He welcomes automation that removes drudgery or gives ordinary workers capabilities previously reserved for the wealthy. He also argues that productivity savings should reach people through lower living costs rather than being erased by inflation. The framework turns those statements into an adoption review: define the burden removed, test whether users gain choices, assess effects on meaningful work, trace who captures the savings, and revisit the decision as evidence changes. It does not establish that AI will produce abundance or unemployment; both speakers repeatedly acknowledge uncertainty.
Origin
Extracted from The Diary of a CEO
Core principles
- 01Technology should increase people's control over their lives
- 02Material convenience does not replace meaning or contribution
- 03Automation gains should lower costs for ordinary users
- 04Uncertain forecasts still permit stable decision principles
How to run it
- 1
Define the human burden
Specify which task, cost, delay, or constraint the technology changes. Identify the people affected directly and indirectly.
Pro tip Separate drudgery removed from capabilities or relationships that users value.
Watch out Efficiency for the buyer may transfer risk or unpaid work to someone else.
- 2
Test agency
Ask whether people gain useful choices, control, and access or become more dependent on a provider or system. Examine consent and the ability to override or leave.
Pro tip Observe what users can actually decide after adoption, not what the product promises.
Watch out Convenience can conceal lost privacy, autonomy, or recourse.
- 3
Assess meaning and contribution
Identify how the change affects opportunities to work, learn, create, care, and contribute. Redesign roles where automation removes tasks but people still need a path to mastery and participation.
Pro tip Ask affected workers which parts of the role provide purpose as well as income.
Watch out No single form of work supplies meaning for everyone.
- 4
Trace the gains
Measure productivity, price, time, and income effects and identify who captures them. Check whether users and workers share in lower costs or higher output.
Pro tip Set a baseline before deployment so claimed savings can be tested.
Watch out A productivity gain does not automatically become a consumer price reduction.
- 5
Protect participation
Add training, role redesign, access, or transition support where the technology would otherwise exclude people from useful contribution. Keep safeguards proportionate to observed and plausible harms.
Pro tip Pilot with the most affected group before broad rollout.
Watch out Principles do not replace legal, technical, or sector-specific safety review.
- 6
Reassess under evidence
Monitor adoption, displacement, cost, agency, and user outcomes. Change the policy when actual effects differ from the original assumptions.
Pro tip Predefine indicators that would pause, expand, or redesign the deployment.
Watch out Confidence in a technology leader's forecast is not evidence of the forecast's timing.
In the wild
Poilievre imagines a minimum-wage worker gaining a personal assistant and automated transport, freeing time otherwise spent on drudgery. He supports that direction only alongside continued opportunities to work, contribute, and find meaning, with productivity gains reflected in lower costs.
→ The example passes the test only if added convenience also preserves agency and participation; the episode does not establish that this future will occur.
A company pilots an AI assistant for repetitive ticket classification. It lets staff override decisions, measures time saved and error rates, retrains workers for customer problem-solving, and shares gains through shorter queues and development time rather than quietly removing every entry-level role.
→ The pilot tests agency, customer benefit, and continued contribution alongside efficiency.
Common mistakes
Equating convenience with empowerment
A faster service can still reduce control, privacy, or opportunities to contribute.
Assuming gains will be shared
Lower production costs do not automatically reach workers or consumers without measurement and deliberate choices.
Planning from one forecast
The speakers stress that no one knows the exact trajectory, so the framework requires review as evidence develops.
Is it for you?
Best for
It is best for product, workplace, or policy choices where automation may change people's autonomy, work, costs, or sense of contribution.
Not ideal for
It is not ideal as a quantitative labor forecast or a substitute for evidence about safety, discrimination, privacy, security, and economic effects.
From the transcript
“I have principles that I would apply as these technologies present themselves”
“ensure that it becomes an enabler of humanity, not a replacement for it”
“the rules around technology should always be geared towards giving people more agency, more meaning, and more control”
From the episode
Pierre Poilievre, The Next Prime Minister of Canada?: The Economy Is About To Collapse!