The AI Inevitables Map
Plan around forces you cannot stop and risks you can still influence
- Difficulty
- Advanced
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 91%
The AI Inevitables Map organizes Gawdat's argument around three recurring propositions: AI development is unlikely to stop because competitors do not trust one another to pause; machines are expected to become substantially more capable; and disruptive or harmful outcomes will occur alongside beneficial ones. He later adds a fourth belief: greater intelligence may favor abundance over scarcity. The practical mechanism is to treat the first forces as planning constraints, distinguish immediate human-led misuse from more speculative autonomous threats, and direct effort toward increasing the probability of beneficial outcomes. This is a scenario-planning model, not a settled forecast. Its value lies in replacing a binary utopia-versus-extinction debate with explicit assumptions, multiple outcomes, and actions that remain available before capabilities exceed effective human control.
Origin
Gawdat says the inevitables structure his book Scary Smart and revisits them in The Diary of a CEO.
Core principles
- 01Competitive distrust makes a universal stop unlikely
- 02Machine capability is expected to continue increasing
- 03Greater capability creates both beneficial and harmful possibilities
- 04Planning should focus on influenceable outcomes rather than certainty
How to run it
- 1
Identify the unstoppable force
State what competitive, technical, or political dynamics make a complete halt implausible.
Pro tip Name the actors and incentives rather than calling progress inevitable by default.
Watch out An asserted inevitability is still an assumption that should be revisited.
- 2
Map the capability curve
Describe how increasing capability could change who can perform a task and at what scale.
Pro tip Use ranges and scenarios instead of unsupported precision.
- 3
Prioritize proximal disruption
Assess job, security, power, and information effects that could arrive before science-fiction scenarios.
Pro tip Separate observed effects from forecasts.
Watch out Do not let dramatic distant risks hide nearer human misuse.
- 4
Branch the outcomes
Build plausible positive and negative paths, noting what would have to happen for each one.
Pro tip Include outcomes where the technology changes society without either saving or destroying it.
- 5
Shift the probabilities
Choose ethical development, governance, education, or market actions that make preferable paths more likely.
Pro tip Assign an owner and an observable leading indicator to each action.
Watch out Do not present influence over an outcome as control over it.
In the wild
Illustrative example: a media company assumes synthetic production will continue improving, maps likely effects on routine editing and audience trust, and separates near-term impersonation risks from speculative autonomous threats. It adopts provenance controls, trains staff to use approved tools, and invests in human-led formats whose value is genuine access and connection.
→ The company prepares for disruption while increasing the usefulness and safety of its own AI adoption.
Common mistakes
Treating a forecast as a fact
Gawdat's timelines, intelligence comparisons, and probabilities are claims made in the interview, not established results.
Planning only for extinction
The episode repeatedly argues that nearer human misuse and economic disruption deserve attention first.
Using inevitability as an excuse
The model treats some trends as constraints but still asks people to influence which outcomes become more likely.
Is it for you?
Best for
Leaders, investors, developers, and policymakers planning around fast-moving AI capabilities.
Not ideal for
Precise forecasting, because the episode provides opinions and scenarios rather than validated probability estimates.
From the transcript
“The first inevitable is AI will happen”
“the second inevitable is they'll become significantly smarter”
“the third inevitable is bad things will happen”
From the episode
EMERGENCY EPISODE: Ex-Google Officer Finally Speaks Out On The Dangers Of AI! - Mo Gawdat