You're guilty until proven innocent. Read that again.

As our systems have evolved—from criminal courts to startups to advertising, and everything in between—I’ve come to observe that we’ve continued to build them on a baseline assumption: good faith by default.

But I’ve also come to believe that in the post-AI world, that assumption collapses.

The more integral GPT has become to my daily thinking, the more reasonable it becomes to shift our baseline assumption: bad faith unless proven otherwise.

And that our systems should evolve to match.

Proposed Principle: “Guilty Until Proven Innocent”

The Rules of the Principle

Each domain sets its baseline differently based on who holds power, who bears risk, and how expensive it is to cheat.

Here’s why the burden of proof flips depending on the context:

1. Murder Conviction (Defendant’s Behavior)

Should good faith be assumed? → ✅ Yes

• Cost of deception: Extremely high (life in prison or death)

• Reward of deception: Extremely low (lying rarely helps; often backfires)

Conclusion:

The system assumes innocence because deception is too risky, and most people have no incentive to lie.

2. Startup Investing (Founder Behavior)

Should good faith be assumed? → ✅ Yes

• Cost of deception: High (lost trust, capital access, reputation)

• Reward of deception: Low (short-term optics, long term collapse)

Conclusion:

Founders are directly exposed. They hold the money. They bear the risk.

The cost of amoral behavior is personal and unrecoverable so good faith can be reasonably assumed.

3. Venture Capital Fundraising (VC Behavior)

Should good faith be assumed? → ❌ No

• Cost of deception: Low (minimal LP oversight, rare consequences)

• Reward of deception: High (management fees, carry, new fund commitments)

Conclusion:

VCs operate with asymmetric leverage and face minimal downside for narrative manipulation especially under the accepted norm that 90%+ of portfolio companies will fail and funds won’t mature for 10 years.

They can offload accountability for their own misjudgments and still raise the next fund so bad faith can be reasonably assumed.

4. Business-Led Advertising (Marketer Behavior)

Should good faith be assumed? → ✅ Yes

• Cost of deception: High (lawsuits, brand damage, job loss)

• Reward of deception: Medium (short-term sales, long-term reputational risk)

Conclusion:

Employees and legal teams are naturally and structurally aligned around compliance.
The system is centralized, monitored, and reputationally accountable so true good faith effort by all involved parties can be assumed unless proven otherwise.

5. Creator-Led Advertising (Influencer Behavior)

Should good faith be assumed? → ❌ No

• Cost of deception: Low (rare enforcement, accountable primarily to “self”)

• Reward of deception: High (virality, more views = more money and more praise , validation and praise from followers)

Conclusion:

Incentives reward dishonesty. Oversight is weak. Accountability is fragmented. Bad faith can be reasonably assumed unless creator is proactively and continuously proving good faith in all dealings.

Kaeya