OpenAI Podcast: Episode 13 – The Thinking Behind Ads in ChatGPT
Date: February 9, 2026
Host: Andrew Mayne
Guest: Asad Awan
Episode Overview
This episode explores OpenAI's decision to introduce ads into ChatGPT, focusing on the reasoning, guiding principles, user experience, and future directions. Andrew Mayne interviews Asad Awan, who provides an insider perspective on how ads will be implemented, how they'll look and feel to users, the controls in place to protect privacy, and how OpenAI is orienting its business model around user trust and broad accessibility.
Key Discussion Points and Insights
1. Why Ads, and Why Now?
- Accessible AGI: OpenAI wants to "bring AGI to all of humanity and to benefit all of humanity" ([00:33]).
- Scaling for All: With over 800 million users, ads are seen as a proven model to fund robust, accessible AI for everyone ([00:33]).
- Balancing Act: Ads help avoid severely limiting the free tier’s usefulness, ensuring broad access without sacrificing quality ([01:16], [08:22]).
- “If the goal is to truly democratize access, I think ADS is a good model.” – Asad ([01:35])
2. Guiding Principles for Ads in ChatGPT
- Separation Between Answers and Ads: Ads are strictly separated from model-generated answers, both visually and technically ([04:01], [04:30]).
- “The answers need to be independent from the ads, both visually but also in how the models are trained and how the system works so that you can always trust the answer.” – Asad ([01:35])
- Model has no awareness of the ad’s content unless the user directly engages ([04:12]–[05:11]).
- Privacy and Control: Sensitive conversations will never trigger ads; user conversations are never shared with advertisers ([01:35]).
- Matching is handled internally and advertisers can’t see user-level data ([01:35]).
- “If you have a sensitive conversation that will never have ads in it, and the conversations are never shared with advertisers.” – Asad ([01:35])
- Transparency: Users have access to information about what data is used for ad targeting and can control what’s shared ([11:12]).
- User Trust as Core Business: OpenAI prioritizes user trust above all else, above even user and advertiser value ([08:57]).
- “Our business model is trust.” – Asad ([05:39])
- Conservative Approach: Ads will only be shown if they’re genuinely relevant and useful; otherwise, none will be shown ([12:57]).
- “If it is not, we'd rather not show you anything.” – Asad ([12:57])
3. Personalization, Controls, and User Experience
- Who Sees Ads? Free and ‘Go’ tier users see ads; Pro, Plus, and Enterprise users do not ([07:38]).
- User Controls:
- Users can restrict or clear data shared for ads ([11:12]).
- Ability to turn off personalization or upgrade to avoid ads ([11:12]).
- “You can clear your data ... don’t use my past chats, if that’s what you care about. Or ... turn off personalization fully.” – Asad ([11:12])
- Visual Design: Ads will be clearly distinct, typically as a bottom-banner, separated visually from model answers ([04:30], [14:43]).
- Frequency of Ads: Rollout will be conservative, with relatively few ads at first ([12:57]).
- Sensitive Content: Advanced systems define and detect sensitivity (health, politics, violence); such topics are excluded from ad placement ([13:44]–[14:36]).
4. Governance, Decision-Making, and Company Culture
- Principles Over Revenue: Structured rubric: user trust > user value > advertiser value > revenue ([08:57]).
- “User trust more than user value, which is then more important than advertiser value, which is more important than revenue.” – Asad ([08:57])
- Company-Wide Input: Hundreds of internal roundtables contributed to the development of ad principles ([08:57]).
- Guardrails: Strict policies guide what constitutes sensitive content and appropriate ad contexts; ongoing governance built in ([16:13]).
5. Ads in Practice and the Advertiser Perspective
- Small Business Access: Lowering the technical/financial barrier for SMEs to run successful, targeted ads, potentially via conversational prompting ([20:38]).
- “The vision has to be where almost as easy as you are prompting ... you could say, my goal is sell these shoes more in Midwest and go.” – Asad ([20:38])
- Agentic Future: Envisions AI as an agent aggregating best deals based on user preference, transforming ad discovery and efficiency ([24:23]).
- “The next version would be can it work behind the scenes and actually aggregate the best discounts and best deals and the best version of the product.” – Asad ([24:23])
6. Addressing Criticism and Industry Perspectives
- Anticipating Skepticism: Recognizes users' reluctance to see ads, promises transparency and offers ad-free upgrade options ([17:34]).
- OpenAI’s Holistic Mission: Other companies may dismiss ads, but OpenAI’s mission to serve all—consumers, enterprises, and diverse business models—requires flexibility ([19:02]).
- “If there is this elitist view that some people get to use it and some don’t ... I think that itself is a pretty big fork in the road in terms of how AI can be valuable to people.” – Asad ([20:00])
Notable Quotes & Memorable Moments
-
On Trust as Core Value:
“Our business model is trust. This is very different than many other scenarios.” – Asad ([05:39]) -
On User Control:
“The users need to believe and understand and control what’s happening.” – Asad ([08:57]) -
On Ad-Model Separation:
“The model doesn’t know whether an ad is there or not. If you ask it, ‘Hey, what’s this ad saying?’ It’ll say, ‘I actually don’t know ...’” – Asad ([04:12]) -
On Advertising for Small Businesses:
“The vision has to be where almost as easy as you are prompting nowadays for questions ... Do you want to spend more money on this? ... It almost becomes an agent for that.” – Asad ([20:38]) -
On Future of Ads:
“The next version would be can it work behind the scenes and actually aggregate the best discounts … Like, for example, if I know that I like ramen ...” – Asad ([24:23]) -
On Responding to "No Ads" Feedback:
“I do believe that they have valid questions and concerns around privacy; it’s on us to do a really good job to earn their trust through better transparency, through better control, through building. That is also delightful.” – Asad ([17:41])
Timestamps for Key Segments
- Why Ads, Why Now? – [00:33]
- Guiding Principles for Ads – [01:35]
- Separation: Answers vs. Ads – [04:01], [04:30]
- Business Model is Trust – [05:39]
- Ad-Free Tiers vs. Ad-Supported Tiers – [07:38]
- Company Decision Process & Prioritization – [08:57]
- Personalization, Controls & User Experience – [11:12]
- Sensitive Content Guardrails – [13:44]
- Visual Design of Ads – [14:43]
- Advertiser Experience, SMBs, and the Agentic Future – [20:38], [24:23]
- Direct Response to Ad Critics – [17:41]
- Competing Industry Perspectives – [19:02], [20:00]
Concluding Thoughts
The episode spotlights OpenAI’s determination to expand access to advanced AI while addressing user trust and privacy. Asad Awan explains the company's measured and user-centric approach to deploying ads, the meaningful controls and transparency offered to users, and how high standards underpin every step—promising an advertising model that could mark a significant break from industry norms. As OpenAI tests these ideas in the real world, it remains committed to adjusting and evolving based on feedback, striving to prove that ads and user trust can coexist in service of making AI truly universal and beneficial.
