OpenAI's comms team has a Sam Altman problem | The San Francisco Standard

By Heather Kelly Technology Editor Published Aug. 13, 2026 at 6:00am Share Share Copy link to this article Email (opens in new tab) Twitter (opens in new tab) Bluesky (opens in new tab) Facebook (opens in new tab) LinkedIn (opens in new tab) Telegram (opens in new tab) WhatsApp (opens in new tab) Reddit (opens in new tab) Close share menu There are $ 0 comments on this story Published Aug. 13, 2026 at 6:00am Being the head of communications for a controversial, billion-dollar AI company is not an easy job under any circumstances. But Silicon Valley comms veteran Aaron Zamost thinks there’s something especially impossible about the position at OpenAI, which has been searching for a chief communications officer since January. Zamost has been working for big-name tech companies for decades, crafting their messages, navigating crises, and handling big personalities. He worked at Google for five years, then at Square for nine as head of communications and public policy, then head of people, reporting directly to CEO Jack Dorsey. Now he runs his own shop while guest lecturing at Berkeley School of Law and organizing events, like the upcoming reunion (opens in new tab) of pioneering tech site ReCode with Kara Swisher. The one thing he’s not doing is trying to be the next head of communications for OpenAI. We spoke with Zamost about the pitfalls of managing a volatile CEO and why OpenAI can’t stop playing defense. ADVERTISEMENT This interview has been edited and condensed for clarity. You’ve said one of the big problems OpenAI is having is that the chief spokesperson for the company is also a widely distrusted figure: Sam Altman. Is that common in the industry? I would argue that Meta has the same problems as OpenAI, which is that the person in charge [CEO Mark Zuckerberg] doesn’t know how to be relatable, and therefore people don’t want to use their product, because they associate that product with the person they don’t like. People don’t use Meta and Instagram happily, and I worry that OpenAI is fumbling an opportunity to get people to want to use their products happily as well, and that has trickle-down effects to the rest of the AI industry. Do you think OpenAI’s main rival, Anthropic, suffers from a similar issue? Is CEO Dario Amodei just more likable? Dario has purposefully been less visible than Sam. I think Anthropic has been mercilessly on message. Whether you agree with it or not, when people think of Anthropic, they understand that it cares about safety, so I think Dario gets more slack than Sam. If you have seen Sam Altman out publicly for the last several months, you have heard him say that his own product is potentially a catastrophic and existential threat to society and humanity, while at the same time messaging, “Use ChatGPT.” I mean, how does that even work? But Anthropic has messaged that danger too, as it did with the paused rollout of its LLM Fable, which it originally said was too powerful to release to the general public. That was everyone else’s perspective, not Anthropic’s. Anthropic isn’t pitching everyone on how horrible they are. Do AI companies sometimes want to make themselves seem more important by making their products seem scary? Is that the marketing tactic? If so, it’s a terrible idea. Apple isn’t making you horrified daily of the fact that they have a digital record that can identify the faces of you and every member of your family. Well, when you put it that way. We generally don’t try to tell people about the worst use cases of our own products. OpenAI and Sam are drawing attention to the horrible things about their product that people are already thinking: You might not have a job in the future. We’ll figure out universal basic income, but that will have to come after we destroy American labor. It just doesn’t make any sense. Sam and Mark may be tech visionaries — although, given the metaverse, that’s debatable for Mark — but they seem constitutionally incapable of acting like normal people. More about OpenAI The AI boom is reshaping San Francisco’s economy. Renters are feeling it first Ideas AI’s new political donor class is already outspending Big Tech’s last one The myth of bartering AI stock for real estate How would you describe OpenAI’s current approach to comms? A lot of OpenAI’s comms and marketing bench came up through defensive disciplines. Primarily, many of them came from Meta, and they are absolute experts at crisis response and policy- and litigation-adjacent work. That’s a real skillset, and it’s valuable when you’re playing defense. But it’s a completely different muscle than telling a positive product story that makes people fall in love with something. And so what they’ve done is build a roster of all defensive players and then wonder why they can’t score any goals. Even if they have fantastic consumer comms people, even the best pitch or article or story you tell in the world gets buried if your CEO is out doing 10 minutes with Tucker Carlson (opens in new tab) , defending himself against murder allegations. Or goes on Jimmy Fallon (opens in new tab) and says he could never have raised a newborn without ChatGPT, even though every human being in the history of civilization has. You cannot build public trust and goodwill with one hand while, on the other hand, the CEO serves up the exact fear and weirdness that make people distrust AI in the first place. What if they really are concerned about the impact of their products and are trying to get ahead of a real threat? How should they address that? What is the comms strategy for “Actually, yeah, we might put you all out of work”? Have thoughts on this story? Start the conversation I’ll be a good comms person and reject the premise of the question. I think they need to start primarily by focusing much more strongly on telling positive stories about what this technology can do that are not vague and future-oriented. “This might change medicine and have fantastic health outcomes from the world” does not persuade me, because it sounds like science fiction, and it’s coming from an industry that spent the better part of a decade telling regular people how important crypto was. So people are skeptical of these unsubstantiated arguments, but there are very real use cases for AI. I think OpenAI in particular needs to act less like Meta and more like Apple or Nike. Let’s say you’re put in charge of comms at OpenAI, which is a job you’ve said you don’t want. What are some changes you would make? I would hire a lot of people from companies with beloved consumer products. The brilliance of the early days of Peloton communications was that their best spokespeople were their customers, which is always the case for the best companies. The iPhone was always about people telling you it was the greatest phone in the world, not the company telling you that. And currently, no one is telling you that ChatGPT is any good except for the people at ChatGPT or at OpenAI. You need to find comms people who know how to activate those feelings and, in partnership with marketing, find customers to tell stories, as opposed to you. Do you want me to answer the question about why I think that job is impossible? Yes, I’d love that. The reason the job is impossible is because Sam will step on that strategy tomorrow with some other crazy thing about how you will pay for AI like we pay for water. Or he’ll get in a pissing match online with Elon [Musk], and it’s just distracting from the most important message that OpenAI should be putting out in the world. I think the reason they haven’t filled the role is because of the fact that it’s impossible, because the head of comms is ultimately the one judged for the communication strategy of the company. And if you cannot control the CEO, everything that Sam does looks like a choice of the head of communications. As I understand it, Sam speaks directly with reporters himself and says what he wants, whenever he wants. And good for him to want to be authentic and not intermediated by his comms team. But then, when he does say the things he says, that would make the head of communications look like an idiot. I think the best people in comms understand that and — let’s be honest — want to manage their own reputations, and therefore the only person they could hire would be someone who would be OK with that dynamic, and that person is probably too junior for the role itself. What’s interesting to me about your perspective on this is you’ve worked with really big, well-known CEO personalities like Jack Dorsey. Is there a secret to wrangling these characters? Executives hate being handled, and the best executives don’t appear handled. So the question, as always, is how do you handle them in a way that seems authentic to who they are, and for me, that means working with the person as they are and taking your wins at the margins. So no, you cannot media-train a person to be unlike who they are. No comms person in Salesforce can convince Marc Benioff not to take an interview on a massage table. [LINK] And I, as a reporter, love that. Of course you do. There are things you have to accept about the way your executives are, and then work with what you have. I think the executives I’ve worked with who are the best at communicating do understand that communications itself is an art, not a science, and not everyone can do it well. The ones who fuck up the most are the ones who think they’re great at it and don’t need any help. You’ll rarely find Tim Cook going head-to-head with somebody on social media. Because he’s a grown-up. I think the comms people at OpenAI are probably doing the best within the system Sam has created. They’re playing a bad hand as well as they can. Can you describe the difference between regular comms and crisis comms for a company like OpenAI? People think of crisis comms primarily through a reactive lens: Bad thing happens; company responds. And that’s how you get statements that are straightforward and seek to make the bad thing less bad. That’s why the default statement on litigation is, “These are baseless claims, and we look forward to rebutting them in court.” A lot of crises require you to show emotional self-awareness. So while the lawsuits about ChatGPT potentially counseling people to harm themselves might be legally baseless, taking that approach in response is crazy. Instead of saying, “Our hearts go out to the families affected,” then attempting to defend yourself on legal terms, the better statement is: “This is horrible, and we are doing everything we can to make sure this never happens again.” Crisis comms is being proactive before the crisis happens, so there’s a positive story being told at the same time as the negative defensive story. Right now, even if OpenAI is doing that work, it is being overwhelmed by what the media wants to cover, which is the shitty, awful things. So it’s all the media’s fault? The big complaint you hear from executives in the Valley is that reporters hate tech. They hate them. They’re jealous. They’re out to get them. Everything’s a hit piece, and I think that misunderstands the supply-and-demand dynamics of newsrooms. We’ve created a virtuous cycle where Silicon Valley’s self-styled elites love to hate on the media, not understanding that the dynamic is economic, not cultural. So I will ask you, what stories about AI would you like to be pitched? It has to be something tangible. We want to see it actually changing somebody’s life and not in such a packaged, precious way. What’s funny is people in journalism aren’t haters. We’re using these tools ourselves. We are fans and critics. But a lot of founders and tech figures want to be in charge of their message, to write a Medium post and skip going through traditional outlets. I think it’s a huge mistake. The credibility of an outlet like Wired writing about your new technology (opens in new tab) gives you 100 times the benefit than an owned channel or friendly podcast would. And if an executive is too blind to see that even small legitimate criticism actually makes the article more credible, then shruggy emoji, right? I used to say to my team when they’d read an article they didn’t like, “Is anything in it factually wrong?” If the answer was yes, then that is a legitimate reason to be upset, and you should fight back with the reporter who worked on it. But ultimately you have to accept a fair shake and reasonable analysis. I find there’s a lot more emotional, instant reactions from companies. The problem, as always, is everyone reads the news, everyone can write, and everyone can talk, and therefore everyone thinks they’re a communications expert. I’m not an engineer, so I can’t comment on the quality of your code. I am not in finance; therefore, I can’t comment on the accuracy of your model. But because communications is the thing everyone can do, executives have very strong opinions on whether you’re doing it well. I often say, when life gives you shitty lemons, you make shitty lemonade. And right now, OpenAI is making shitty lemonade, and it’s the best that they can do given the system they’re in and the current environment around AI — which is something their own boss created. Copy link to this article (opens in new tab) (opens in new tab) (opens in new tab) (opens in new tab)