A practical guide with copy-and-paste AI prompts for researching candidates, checking political claims, comparing promises, following the money, and making your own evidence-based decision.
The 2026 local elections are approaching across British Columbia.
On October 17, 2026, communities across BC will elect local representatives including mayors, councillors, regional district directors, school trustees and other local officials. Candidate nominations run from September 1 to September 11, 2026, meaning candidate information researched before nominations close may still be incomplete.
Local elections can affect things we interact with every day: roads, development, property taxation, recreation, local infrastructure, policing decisions, water and sewer services, planning and many other community services.
But making an informed decision can be difficult.
A voter may encounter:
- candidate websites
- campaign flyers
- Facebook posts
- TikTok and YouTube videos
- debates
- news articles
- endorsements
- attack ads
- statistics without context
- screenshots with no source
- AI-generated content
- friends and family giving opinions
The problem isn't necessarily a lack of information.
The problem is figuring out what information deserves your trust.
Artificial intelligence can help.
But there is an important distinction:
Don't ask AI to decide who you should vote for. Ask AI to help you understand what you're voting for.
Used properly, AI can act more like a research assistant: collecting records, comparing candidates under the same criteria, checking claims against primary sources, identifying uncertainty, and helping you examine whether a campaign promise is realistically achievable.
Used poorly, AI can simply reinforce your existing opinion—or confidently repeat incorrect information.
Elections Canada specifically recommends checking the original source of information, checking whether sources are legitimate, comparing information across sources, and making sure information is current. It also warns Canadians to be cautious about inaccurate information circulating online.
This guide gives you prompts you can copy directly into an AI assistant and use yourself.
First: Understand What AI Can and Cannot Do
AI should not be treated as an election authority.
It can make mistakes.
It can misunderstand legislation.
It can confuse provincial, federal and municipal responsibilities.
It can repeat an incorrect news story.
It can misinterpret a candidate's position.
And it can sometimes produce a confident answer when the available evidence does not justify that confidence.
For election research, a good AI response should therefore do three things:
1. Show its evidence.
2. Separate fact from interpretation.
3. Clearly admit when something cannot be established.
The objective is not to eliminate every possible form of bias. No research system can guarantee that.
The objective is to create a process that makes bias, uncertainty and weak evidence harder to hide.
Start Every Election Search With This Prompt
Before asking AI about individual candidates, give it rules.
Copy this entire prompt.
```text You are acting as a non-partisan voter research assistant.
Your purpose is NOT to persuade me to vote for or against any candidate.
RESEARCH RULES:
1. Apply exactly the same research standard to every candidate.
2. Search current information. Do not rely only on your existing knowledge.
3. Prefer primary sources whenever possible, including:
- Elections BC - BC Government - BC Laws - the municipality - regional district - school district - official budgets - audited financial statements - council or board minutes - recorded votes - official meeting recordings - Statistics Canada - official public datasets
4. Candidate websites, social-media accounts, campaign literature and speeches may establish WHAT A CANDIDATE SAYS.
They do not automatically establish that what the candidate says is true.
5. Use reputable journalism as secondary evidence.
6. For important information, classify the statement as:
VERIFIED FACT CANDIDATE CLAIM ANALYSIS / INFERENCE OPINION DISPUTED UNKNOWN / INSUFFICIENT EVIDENCE
7. Cite the source and date for every important factual claim.
8. Search for credible evidence that contradicts major claims, not only evidence supporting them.
9. Do not infer motives, personality, ideology or intentions unless supported by evidence.
10. Apply the same scrutiny to incumbents and challengers.
11. Determine whether campaign promises actually fall within the authority of the office being sought.
12. Clearly distinguish between:
- what the candidate promises - what the candidate can directly control - what requires council or board approval - what requires provincial involvement - what requires federal involvement - what is outside the jurisdiction
13. Do not determine the "best candidate" unless I first provide my priorities and weighting criteria.
14. When evidence is unavailable, say:
INSUFFICIENT EVIDENCE
Do not guess.
15. When credible sources disagree, explain the disagreement and present the evidence supporting the different conclusions.
16. Do not create false balance.
A claim supported by strong evidence should not receive the same evidentiary weight as an unsupported allegation.
At the end of your research provide:
WHAT WE KNOW WHAT THE CANDIDATES CLAIM WHAT THE EVIDENCE SUPPORTS WHAT REMAINS UNCERTAIN QUESTIONS VOTERS SHOULD ASK PRIMARY SOURCES SECONDARY SOURCES ```
This simple step can substantially improve the usefulness of everything that follows.
Prompt 1: What Can This Politician Actually Control?
Before evaluating a promise, understand the job.
This is especially important in local elections.
Municipal decisions are generally made collectively by council. The Province of BC explains that council members make collective decisions through majority votes; mayors and councillors have defined responsibilities rather than unlimited individual authority.
So if a candidate says:
“I will fix housing.”
Or:
“I will reduce crime.”
Or:
“I will cut your taxes.”
Your first question should be:
Can this person actually do that?
Copy:
```text I am voting in the 2026 election in:
[CITY / REGIONAL DISTRICT], British Columbia.
I am researching the position of:
[MAYOR / COUNCILLOR / SCHOOL TRUSTEE / REGIONAL DISTRICT DIRECTOR]
Before discussing candidates, explain what this elected position can actually control.
Research:
- legal authority
- budget authority
- taxation authority
- land-use authority
- housing responsibility
- infrastructure responsibility
- policing/public safety responsibility
- transportation responsibility
- environmental responsibility
- healthcare responsibility
- education responsibility
- economic-development responsibility
For each issue classify the office's authority as:
DIRECT CONTROL SHARED CONTROL INFLUENCE ONLY PROVINCIAL RESPONSIBILITY FEDERAL RESPONSIBILITY OUTSIDE THE OFFICE'S AUTHORITY
Use BC legislation, the Province of BC and the relevant local government as primary sources.
Explain the difference between:
1. what this individual politician can do alone 2. what requires a council or board vote 3. what municipal staff administer 4. what requires another level of government
Finally identify common campaign promises for this office that could be misleading if presented as something one elected official can accomplish alone. ```
This prompt can quickly turn vague campaign language into a much more useful question:
Who actually has the authority to make this happen?
Prompt 2: Compare Every Candidate Using the Same Standard
Bias can creep into research when we scrutinize one candidate intensely while giving another candidate the benefit of the doubt.
A better approach is to use the same template for everyone.
Copy:
```text Research every confirmed candidate currently running for:
[OFFICE]
in:
[CITY / COMMUNITY], British Columbia.
IMPORTANT:
Verify that the candidate list is current.
If nominations have not closed, clearly state:
CANDIDATE LIST MAY NOT YET BE FINAL.
Do not rank the candidates.
Create an identical profile for every candidate.
Use these exact categories:
1. Professional experience relevant to the position
2. Previous elected or public-service experience
3. Major stated priorities
4. Housing and development
5. Property taxes
6. Government spending
7. Infrastructure
8. Public safety
9. Transportation
10. Local businesses and economic development
11. Environment
12. Major capital projects
13. Transparency and governance
14. Specific measurable promises
15. Estimated costs of major promises where available
16. Proposed funding mechanisms
17. Previous voting record if the candidate is an incumbent
18. Attendance or participation records where reliably available
19. Endorsements
20. Elector organization affiliation
21. Publicly available campaign-finance information
22. Important unanswered questions
For every candidate clearly separate:
VERIFIED RECORD STATED POSITION CAMPAIGN PROMISE INDEPENDENT EVIDENCE UNKNOWN
Apply exactly the same evidentiary standard to every candidate.
Do not describe candidates using political adjectives such as:
radical moderate extreme progressive conservative controversial popular
unless the description is objectively defined, directly relevant and supported by evidence.
Cite every important factual claim.
End with a side-by-side comparison table.
Do NOT recommend a candidate. ```
The purpose is not to make all candidates appear equally good.
It is to make sure they are all evaluated equally carefully.
Prompt 3: Try a Blind Candidate Comparison
Humans have biases too.
Maybe you recognize someone's name.
Maybe you dislike a political organization.
Maybe a friend told you Candidate B is terrible.
Maybe you simply like the appearance or personality of one candidate.
One interesting experiment is to temporarily hide the identities.
Copy:
```text Compare the major candidates for:
[OFFICE AND COMMUNITY]
using a blind comparison.
First research the candidates.
Then temporarily replace their names with:
Candidate A Candidate B Candidate C Candidate D
Do NOT initially reveal:
- candidate names
- elector organization
- endorsements
- political labels
- photographs
- slogans
Compare only:
- policies
- measurable commitments
- costs
- proposed funding
- feasibility
- jurisdiction
- experience
- previous results
- voting record
- evidence supporting their claims
- evidence contradicting their claims
Do NOT reveal which candidate is which yet.
Ask me to identify my priorities.
Only after I finish evaluating the policies should you reveal the candidate identities. ```
It is not a perfect scientific experiment.
But it can be a useful way to confront your own assumptions.
Prompt 4: Can This Campaign Promise Actually Be Delivered?
Campaign promises are easy.
Implementation is harder.
Use this whenever you hear something that sounds impressive.
```text Analyze this campaign promise:
"[PASTE THE EXACT PROMISE]"
Candidate:
[CANDIDATE]
Office:
[OFFICE]
Community:
[LOCATION]
Do not evaluate the candidate's personality.
Evaluate only the proposal.
Determine:
1. What exactly is being promised?
2. Is the promise measurable?
3. Does this elected office have legal authority to do it?
4. Can this politician accomplish it individually?
5. Would a majority council or board vote be required?
6. Would provincial approval or cooperation be required?
7. Would federal approval or cooperation be required?
8. What would implementation likely cost?
9. Has the candidate provided a cost estimate?
10. Has the candidate explained how it would be funded?
11. Could implementation affect taxes, fees, borrowing or other services?
12. What is a realistic implementation timeline?
13. What existing bylaws, plans, contracts or legislation could affect implementation?
14. What are the largest practical barriers?
15. Have similar Canadian communities attempted something similar?
16. What happened?
17. What evidence supports the candidate's claim?
18. What credible evidence challenges it?
Rate ONLY THE PROMISE:
CLARITY: High / Medium / Low
JURISDICTION: Direct / Shared / Limited / Outside Authority
EVIDENCE: Strong / Moderate / Weak / Unknown
COST CERTAINTY: High / Medium / Low
IMPLEMENTATION DIFFICULTY: Low / Medium / High / Cannot Determine
FEASIBILITY: High / Medium / Low / Cannot Determine
Explain every rating.
Cite the evidence. ```
Notice that the prompt doesn't ask:
“Is this politician lying?”
It asks:
“What does the evidence allow us to conclude?”
That's a much safer and more useful question.
Prompt 5: Audit an Incumbent's Record
An incumbent has something a challenger usually doesn't:
a governing record.
That gives voters information that goes beyond campaign promises.
Copy:
```text Research the public record of:
[CANDIDATE]
who previously served as:
[OFFICE]
in:
[COMMUNITY]
Audit their most recent term.
Do NOT evaluate personality.
Find:
- promises from their previous campaign
- council or board motions they introduced
- recorded votes
- major budgets they supported
- major budgets they opposed
- property-tax decisions
- borrowing decisions
- major capital expenditures
- housing and development votes
- infrastructure decisions
- public-safety decisions
- significant policies supported
- significant policies opposed
- attendance where official data exists
- major measurable outcomes
For previous campaign promises classify each as:
FULFILLED MOSTLY FULFILLED PARTIALLY FULFILLED NOT FULFILLED BLOCKED BY COUNCIL / OTHER GOVERNMENT NO LONGER APPLICABLE INSUFFICIENT INFORMATION
For each classification provide evidence.
IMPORTANT:
Do not assume that something happening during the politician's term means the politician caused it.
Separate:
CORRELATION
from:
EVIDENCE OF CAUSATION
Where an outcome resulted from a collective council decision, clearly say so.
Where provincial, federal or economic conditions contributed, identify them.
Finish with:
PROMISES VS RESULTS MAJOR VOTING PATTERNS MEASURABLE ACCOMPLISHMENTS UNRESOLVED ISSUES WHAT CANNOT FAIRLY BE ATTRIBUTED TO THIS PERSON ```
That final section is important.
Giving politicians credit for everything good—or blame for everything bad—that happened while they were in office is not good analysis.
Prompt 6: Follow the Political Money—Carefully
Elections BC maintains the Financial Reports and Political Contributions System, a searchable database containing financial reports and contributions involving provincial and local candidates, elector organizations and other regulated political participants.
Registered elector organizations—sometimes described as local political parties—can endorse candidates, receive campaign contributions and incur election expenses under BC's local-election rules.
Financial research can provide useful information.
But be careful.
A lawful contribution is evidence that a contribution occurred.
It is not automatically evidence of corruption, control or a political favour.
Copy:
```text Research publicly available political-finance information for:
[CANDIDATE / ELECTOR ORGANIZATION]
Use Elections BC as the primary source.
Identify:
- available campaign-finance reports
- total reported contributions where available
- publicly disclosed contribution information
- campaign expenses
- loans or transfers where applicable
- elector organization relationships
- previous campaign-finance history
- relevant third-party advertising information
Clearly distinguish between:
CURRENT 2026 ELECTION INFORMATION
and
HISTORICAL ELECTION INFORMATION.
Do NOT imply corruption, influence, conflict of interest or wrongdoing merely because a lawful political contribution exists.
Separate:
FACT: A contribution or relationship exists.
from:
UNPROVEN INFERENCE: The contribution influenced a particular decision.
Only make claims about influence, conflicts or wrongdoing when credible evidence supports them.
Provide direct primary sources. ```
That distinction matters.
Money can be worth investigating.
But the existence of money alone doesn't prove motive.
Prompt 7: Fact-Check a Political Claim
Someone posts:
“Taxes increased 40% because of this mayor.”
Another says:
“Our city approved more housing than anywhere else in BC.”
Another says:
“Crime has doubled.”
Instead of immediately believing—or rejecting—the claim, break it apart.
Copy:
```text Fact-check this political statement:
"[PASTE CLAIM]"
Speaker/source:
[CANDIDATE / AD / POST / ARTICLE / VIDEO]
Do not begin by assuming it is true or false.
Break the statement into individually testable claims.
For EACH claim provide:
CLAIM:
VERDICT:
- Supported
- Mostly Supported
- Misleading Without Context
- Unsupported
- Contradicted
- Opinion / Not Fact-Checkable
- Insufficient Evidence
BEST PRIMARY EVIDENCE:
CONTRARY EVIDENCE:
MISSING CONTEXT:
GEOGRAPHIC AREA:
TIME PERIOD:
SOURCE DATE:
If statistics are involved, locate the original data.
Check:
- numerator
- denominator
- measurement period
- geographic boundary
- whether numbers are nominal or inflation-adjusted
- whether totals or percentages are being compared
- whether the statistic measures what the speaker claims it measures
Look for evidence supporting AND contradicting the statement.
Do not evaluate the candidate overall.
Evaluate only the claim.
Provide citations. ```
Elections Canada's digital-literacy guidance similarly recommends returning to the original story, checking sources, comparing multiple sources and ensuring the information is current.
Prompt 8: Start With the Community, Not the Candidates
There is another way to research an election:
Don't start with politicians at all.
Start with your community.
What has actually changed?
Copy:
```text Create a data-driven report on:
[CITY / COMMUNITY], British Columbia
before the 2026 local election.
Where reliable data exists, examine approximately the last 5-10 years.
Research:
- population
- population growth
- housing starts
- housing completions
- rental prices
- home prices
- homelessness
- property taxes
- municipal revenue
- municipal expenditures
- municipal debt
- infrastructure spending
- police spending
- fire/rescue spending
- reported crime statistics
- business licences
- employment
- major developments
- transit
- road infrastructure
- parks and recreation
- water
- sewer
- major capital projects
For every useful metric show:
EARLIEST COMPARABLE DATA
2022 LOCAL ELECTION PERIOD
LATEST AVAILABLE DATA
ABSOLUTE CHANGE
PERCENTAGE CHANGE
INFLATION-ADJUSTED CHANGE where appropriate
SOURCE
Then identify the 5-10 largest measurable changes affecting this community.
IMPORTANT:
Do NOT automatically attribute changes to a mayor or council.
For each major trend distinguish between possible influences from:
LOCAL GOVERNMENT PROVINCIAL GOVERNMENT FEDERAL GOVERNMENT POPULATION GROWTH INTEREST RATES INFLATION PRIVATE MARKET OTHER FACTORS
Only make claims about causation where evidence supports them. ```
Now you have something incredibly useful:
a baseline.
Instead of asking:
“Which candidate sounds better?”
You can ask:
“Which candidate's plan actually addresses the problems supported by the data?”
Prompt 9: Ask for the Strongest Case For AND Against a Policy
One weakness of political conversations is that people often consume only arguments supporting what they already believe.
You can intentionally ask AI to challenge you.
Copy:
```text Analyze this proposal:
"[POLICY OR CAMPAIGN PROPOSAL]"
First present:
THE STRONGEST EVIDENCE-BASED ARGUMENT FOR IT.
Then present:
THE STRONGEST EVIDENCE-BASED ARGUMENT AGAINST IT.
For both sides use credible evidence.
Then analyze:
WHO WOULD LIKELY BENEFIT?
WHO WOULD LIKELY PAY?
WHO COULD BE NEGATIVELY AFFECTED?
ESTIMATED PUBLIC COST
ESTIMATED PRIVATE COST
SHORT-TERM EFFECTS
LONG-TERM EFFECTS
POSSIBLE UNINTENDED CONSEQUENCES
IMPLEMENTATION CHALLENGES
SUCCESSFUL EXAMPLES
UNSUCCESSFUL EXAMPLES
QUALITY OF AVAILABLE EVIDENCE
Finally answer:
WHAT EVIDENCE WOULD WE NEED IN 2-5 YEARS TO DETERMINE WHETHER THIS POLICY ACTUALLY WORKED?
Do not artificially force both sides to appear equally strong.
If the evidence strongly favours one conclusion, explain why. ```
That last instruction is essential.
Neutrality does not mean pretending unsupported claims are equivalent to well-established evidence.
Prompt 10: Build Your Personal Voting Matrix
Only now should you start asking:
Which candidate fits what I care about?
Importantly, your priorities should come first.
Not the candidate.
Copy:
```text Help me build a personal voting decision matrix for:
[2026 ELECTION / LOCATION / OFFICE]
Do NOT ask my political affiliation.
Do NOT show candidate scores yet.
First ask me to distribute exactly 100 points across the issues that matter to me.
Possible categories:
Housing Property taxes Government spending Infrastructure Public safety Transportation Environment Local business Economic development Growth and development Parks and recreation Education Transparency Government efficiency Long-term planning Other
My total must equal 100.
LOCK MY WEIGHTS BEFORE EVALUATING THE CANDIDATES.
Then research every candidate using the same sources and criteria.
Score their documented positions and records against MY priorities.
For every score show:
EVIDENCE
CONFIDENCE LEVEL
IMPORTANT UNCERTAINTY
Do not award points when reliable evidence is unavailable.
Instead mark:
INSUFFICIENT EVIDENCE.
Then conduct sensitivity analysis.
Test whether my candidate alignment changes if each of my top three priorities increases or decreases by 20%.
Finally show:
CLOSEST ALIGNMENT WITH MY PRIORITIES
SECOND-CLOSEST ALIGNMENT
AREAS WHERE I AGREE
AREAS WHERE I DISAGREE
MAJOR TRADEOFFS
MAJOR UNCERTAINTIES
PROMISES WITH LOW FEASIBILITY
Do NOT say:
"You should vote for..."
Instead explain how the candidates align with the criteria I selected. ```
This is a powerful distinction.
The AI doesn't choose your values.
You do.
The AI helps investigate the evidence.
An Even Better Method: Separate Evidence Into Levels
When AI gives you information, not every source should receive equal weight.
You can add this prompt:
```text For every important piece of information assign an evidence level:
LEVEL A — PRIMARY EVIDENCE
Examples: BC Laws Elections BC municipal records official council minutes recorded votes audited financial statements official budgets Statistics Canada official government datasets
LEVEL B — DIRECT CANDIDATE EVIDENCE
Candidate website candidate platform candidate speech candidate interview candidate social-media account
This proves what the candidate said or promised. It does NOT necessarily prove that the underlying claim is true.
LEVEL C — INDEPENDENT SECONDARY EVIDENCE
Established journalism academic research professional analysis credible non-government datasets
LEVEL D — UNVERIFIED OR LOW-CONFIDENCE INFORMATION
anonymous social-media posts Reddit comments unsourced graphics screenshots hearsay unverified blogs viral videos
Label every major claim with its evidence level.
When Level A evidence conflicts with Level D evidence, do not treat them as equally reliable. ```
Now a voter doesn't just see information.
They see how strong the information is.
Ask AI to Show Its Confidence
Add another requirement:
```text For every important conclusion show:
CONFIDENCE: HIGH / MEDIUM / LOW
HIGH: Multiple strong primary sources support the conclusion.
MEDIUM: Evidence exists but is incomplete or partially dependent on secondary sources.
LOW: Evidence is limited, disputed or unclear.
UNKNOWN: There is not enough reliable evidence to make the determination.
Explain briefly why each confidence rating was assigned. ```
This helps reduce one of AI's biggest problems:
sounding certain when it shouldn't be.
Ask AI to Search Against Its Own Conclusion
This is one of the best anti-bias prompts you can use.
```text You have reached a preliminary conclusion.
Before finalizing it, actively attempt to prove your own conclusion wrong.
Search for:
- contradictory government data
- contradictory voting records
- alternative explanations
- corrections
- more recent information
- methodological problems
- missing context
- evidence from credible sources that supports the opposite
conclusion
Then provide:
ORIGINAL CONCLUSION
STRONGEST CONTRADICTORY EVIDENCE
WHETHER THE CONCLUSION CHANGED
FINAL CONCLUSION
CONFIDENCE LEVEL ```
Think of it as asking the AI to be its own skeptical reviewer.
Important: Know Who Runs BC Local Elections
BC local elections are different from provincial elections.
According to Elections BC, each local government is responsible for administering its own local election, including voting, ballot counting, reporting results and accepting candidate nominations. Elections BC administers campaign-financing and election-advertising rules under the Local Elections Campaign Financing Act.
That means your research may need to involve several official sources.
For example:
Election administration and local candidate information: Your municipality, regional district or school district.
Campaign financing and elector organizations: Elections BC.
Municipal authority and legislation: Province of British Columbia and BC Laws.
Population and economic statistics: Statistics Canada and government datasets.
Past municipal decisions: Council minutes, agendas, bylaws, budgets and recorded votes.
Before September 11, 2026: Add This Warning
At the time this guide was prepared, candidate nominations for the 2026 General Local Elections had not yet closed.
Elections BC lists:
Nomination period begins: September 1, 2026 Nominations close: September 11, 2026 Campaign period begins: September 19, 2026 General Voting Day: October 17, 2026
So before September 11, add this to any candidate research prompt:
```text IMPORTANT:
The 2026 BC local-election nomination period may still be open.
Verify the current date.
Verify candidate information with the relevant local government.
If nominations have not closed, clearly state:
THIS CANDIDATE LIST IS NOT YET FINAL.
Do not describe the list as complete. ```
A Simple Rule for Political Information
When you see an election claim, ask five questions:
Who said it?
What's the original source?
When was the information produced?
Does independent evidence support it?
What evidence would prove it wrong?
That final question is especially powerful.
If there is no possible evidence that could change someone's conclusion, you probably aren't looking at analysis anymore.
You're looking at belief.
AI Should Help You Think—Not Think for You
The goal of using AI during an election should not be:
“Tell me who to vote for.”
A better question is:
“Help me understand the evidence so I can decide for myself.”
That changes the entire interaction.
Instead of asking AI for political opinions, you're asking it to help organize:
- legislation
- budgets
- voting records
- campaign platforms
- public statements
- statistics
- financial disclosures
- government responsibilities
- conflicting evidence
- uncertainty
You still decide what matters.
You still decide which trade-offs you're comfortable with.
You still cast the vote.
The Ideal AI Election Research Process
A strong voter-research workflow could look like this:
Step 1 — Understand the office
What can this person actually control?
↓
Step 2 — Understand your community
What problems are actually supported by data?
↓
Step 3 — Research every candidate equally
What have they done, said and promised?
↓
Step 4 — Examine individual promises
Are they realistic, affordable and within jurisdiction?
↓
Step 5 — Investigate incumbents
What happened compared with what they previously promised?
↓
Step 6 — Follow campaign financing
What financial relationships are publicly documented?
↓
Step 7 — Fact-check important claims
What does the original evidence actually say?
↓
Step 8 — Challenge both sides
What is the strongest evidence against your current opinion?
↓
Step 9 — Define your priorities
What actually matters to you?
↓
Step 10 — Compare
Which candidate most closely aligns with those priorities according to the available evidence?
One Master Prompt
If you don't want to use all of the individual prompts above, save this one.
```text Act as my non-partisan research assistant for the 2026 British Columbia local elections.
I live in:
[COMMUNITY]
I am researching:
[OFFICE]
Your job is NOT to tell me who to vote for.
Your job is to help me make my own informed decision.
STEP 1
Explain what this office can and cannot control.
STEP 2
Research the major measurable issues facing my community using official data.
STEP 3
Find the current confirmed candidates.
Warn me if nominations are still open or the candidate list may change.
STEP 4
Research every candidate using exactly the same criteria.
STEP 5
Separate:
VERIFIED FACTS CANDIDATE CLAIMS CAMPAIGN PROMISES ANALYSIS OPINIONS DISPUTED CLAIMS UNKNOWN INFORMATION
STEP 6
For major promises analyze:
jurisdiction cost funding timeline feasibility evidence barriers
STEP 7
For incumbents compare previous promises with actual votes, decisions and results.
STEP 8
Review relevant Elections BC campaign-finance information.
Do not imply wrongdoing without evidence.
STEP 9
Identify the strongest evidence supporting and challenging each candidate's major claims.
STEP 10
Ask me to allocate 100 points across my personal priorities.
Lock those priorities before scoring candidates.
STEP 11
Compare candidate alignment against MY priorities.
STEP 12
Perform a sensitivity analysis to determine whether small changes to my priorities significantly change the result.
For every important factual statement:
CITE THE SOURCE.
Prefer primary sources.
Show the publication or reporting date.
Show confidence:
HIGH MEDIUM LOW UNKNOWN
If reliable evidence does not exist, write:
INSUFFICIENT EVIDENCE.
Never invent missing information.
Never tell me:
"You should vote for Candidate X."
Instead give me the evidence, uncertainty and trade-offs necessary for me to make that decision myself. ```
Final Thought
AI could become extremely useful during elections.
Not because it knows who the “right” candidate is.
It doesn't.
But because, when properly instructed, it can help ordinary voters process an enormous amount of information that would otherwise take hours or days to research.
The opportunity isn't to replace human judgment.
It is to give human judgment better information.
So instead of asking:
Who should I vote for?
Try asking:
What do I need to know before I vote?
That is a much better place to start.
Official Sources Used for This Guide
This article was prepared using current information from Elections BC, the Province of British Columbia, BC Laws, and Elections Canada. Elections BC confirms October 17, 2026 as General Voting Day and provides the official local-election calendar.
The Province of BC provides guidance on the responsibilities of mayors, councillors, municipal councils and other locally elected officials.
Elections BC operates the Financial Reports and Political Contributions System for researching political financial reports and contributions.
Elections Canada provides digital-literacy guidance recommending that Canadians verify original sources, compare information across sources and ensure information is current when encountering potentially inaccurate election information online.
Updated: August 16, 2026
Important: This article is intended to help voters research election information. It does not endorse or oppose any candidate, elector organization or political position.