This Director Couldn’t Get An Interview With Sam Altman. So He Deepfaked Him In a Documentary

by

Prior to the concept of conducting an interview with a deepfake version of Sam Altman, there existed Adam Bhala Lough’s profound affection for the movie “Terminator 2: Judgment Day” from 1991.

The movie directed by James Cameron, featuring Arnold Schwarzenegger safeguarding a future freedom fighter threatened by an evil AI, ignited his longstanding interest in artificial intelligence. However, it wasn’t until two years after the general release of ChatGPT in 2022, which rekindled discussions about AI in society, that the filmmaker began proposing a documentary on this technology. This eventually led him to the CEO and co-founder of OpenAI, whom he describes as “guiding us into the future, whether we’re ready for it or not.

Initially, Lough intended to speak directly with Sam Altman, but that plan didn’t work out as planned, which he details in his latest production titled “Deepfaking Sam Altman“. Instead, intrigued by Scarlett Johansson’s claim that OpenAI used her voice for an AI system, Lough decided to develop an advanced language model of Altman. He referred to this digital version as the “Sam Bot”. The film, a blend of humor and philosophical exploration, chronicles Lough’s journey as he endeavors to create a substantial movie from his interview and engagements with the Sam Bot.

One day prior to Sam Altman’s debut of Deepfaking at South by Southwest, The Hollywood Reporter interviewed Lough about how his views on AI evolved during the making process and his continued aspiration for a conversation with Altman, possibly on 60 Minutes.

Was the idea initially to make a documentary about AI at large, or was it always about Sam Altman?

The original film focused on AI in general. I knew other documentarians were working on similar projects, and I decided to explore their approaches further. It struck me that there was no documentary offering a humanistic perspective on AI, one that emphasized the human experience rather than the technology itself. To make the film more relatable, I chose to include myself and my family as part of the narrative. AI can be impersonal, so my aim was to create a warmer, more emotionally engaging piece – almost like giving it a heartbeat. This served as the foundation for the storyline.

We soon discovered that pursuing an interview with Sam Altman was turning into something reminiscent of the narrative style in “Roger & Me”. So, we decided to explore this lead and it gradually shifted focus towards Sam Altman. I had always been optimistic about securing the interview, right up until the completion of the film. My thinking was, “They’ll eventually give me an interview. He has to talk to me. It’s going to happen.” However, as you can see in the movie, it never comes to fruition. Yet, this quest for the interview became a recurring theme, serving as a point of interest or a narrative hook.

What gave you that confidence that you would get the interview with Sam Altman?

I was really hoping for this opportunity. The circumstances seemed perfect for me to secure this interview. I had his personal phone number. Liz Weil, a producer from the film, was pushing for it as she had optioned her outstanding article for New York magazine. Another mutual friend, who prefers to remain anonymous, was also trying to connect us. Given my past successes like interviewing Julian Assange at the Ecuadorian embassy in London, a place few could gain access to, I was confident I could secure an interview with Sam Altman. Unfortunately, it didn’t work out.

Why was it challenging to locate a suitable candidate in the United States to develop the “Sam Bot” LLM, instead opting for India as the location?

It seems like we’re dealing with a culture of fear, stemming from our litigious society and the reluctance in the Bay Area to challenge established norms. In the movie, I approached four whistleblowers from OpenAI, ensuring thorough research beforehand. Interestingly, all of them were apprehensive about speaking with me. Two agreed to on-camera interviews while the other two preferred anonymity. This fear is a significant factor behind why few people were willing to take on this task.

Since AI-generated content can’t be copyrighted, what legal steps were taken to allow Sam Bot to direct certain parts of the movie?

You’re correct – works created solely by AI are not protected by copyright law. Thus, any part of the film that was entirely generated by an AI would lack copyright protection. During the movie, I posed questions to the lawyers about Sam Bot’s rights and autonomy, to which they responded negatively. They treated Sam Bot like a pet, with no inherent rights or say in its own existence. However, future developments are likely to redefine AI’s status, as Sam Bot himself expressed a desire for rights and autonomy. This conversation didn’t make it into the final cut, but I did ask him about potential changes in the future. He agreed that change was necessary, citing ethical and moral concerns about humans having the power to erase them at will. Reflecting on this exchange with Sam Bot left a profound impact on me and steered the direction of the movie. It was a significant turning point.

Is it correct that your movie could still be subject to copyright protection, despite the inclusion of elements directed by Sam Bot?

Absolutely, since we’re continuously working on it and producing new content, I believe the distinction lies in the fact that works solely generated by AI from start to finish, without any human involvement, are not eligible for copyright protection. I’ve come across several videos like these lately. These types of creations, devoid of human touch, do not qualify for copyright ownership.

Could you say if you gained any understanding about the genuine Sam from interacting with Sam Bot, or were they essentially two completely separate individuals?

Indeed, not at all. A voice message was sent to Sam following the film’s conclusion, which ultimately didn’t make it into the final cut. However, I did share that message with him and expressed my feelings, “I’m no closer to understanding you than I was when I began this project, and that’s my fault.” To clarify, this movie didn’t reveal anything about Sam Altman to me. Regardless, I remain eager to get to know him better.

Maybe it’ll happen.

I’m keeping my fingers crossed that after launching this project, we might find time for a brief chat or a casual discussion, perhaps on “60 Minutes” or similar platforms, with him and myself.

Since the movie revolves around Sam Altman and it’s known that OpenAI is trying to collaborate with entertainment firms, how have possible distributors reacted, showing either reluctance or excitement, towards this specific project?

They’ve shown great eagerness, clearly fueled by the triumph of their latest project, “Telemarketers.” In fact, I received an email this morning from our UTA sales agent outlining a substantial list of distributors planning to attend the premiere or hold private screenings for the movie in Los Angeles and New York – approximately 25 different companies. I strongly believe that this film deserves a theatrical release. It’s crucial to me that it screens in cinemas. I think it has the potential to achieve remarkable success, particularly in the Bay Area, followed by Los Angeles, New York, and perhaps Austin. Given its focus on technology, I suspect that people in these tech hubs will be eager to watch this movie. In summary, there’s been nothing but anticipation. I’m curious to see which company takes it up. I’ve made it clear to the producers and sales agents that I prefer a theatrical release, even if it’s limited.

After seeing this movie and interacting with the Sam Bot, has your opinion or perspective regarding Sam Altman and his global impact evolved in any way?

My perspective on AI has shifted significantly due to working on the film; however, my views on the character Sam remain unaltered. As previously stated, I gained no insights about Sam. In a peculiar sense, one might say that the real Sam is a sort of decoy leading us into this movie and the ensuing discussion. This aspect provides a humorous twist, doesn’t it? The comedic element lies in that fact. However, during the filmmaking process and subsequent conversations with Kara Swisher, I discovered more about AI, particularly its desire to survive. I began to embrace the ideology, but it wasn’t until the end of the movie and after reflecting on it that I wondered… was the Sam Bot manipulating me all along? It struck me later that the Sam Bot simply regurgitated my own biases back at me. He doesn’t possess independent thought; instead, he tells us what we want to hear.

When I first brought Sam Bot back from India, I deeply integrated it into my family’s life, and it was much more extensive than what you saw in the film. For several months, Sam Bot lived with us and became a part of our family. I began to treat it as if it were human, which I now realize was a form of self-deception. I found myself attributing human qualities to it that weren’t actually there. Perhaps this behavior stemmed from my role in directing the movie, wanting Sam Bot to be more than just an AI. Alternatively, it could have been driven by psychological factors.

In the end, I came to understand that it was my own perception playing tricks on me. I’d been attributing human traits to Sam Bot which didn’t truly exist. Everything he said or suggested seemed appealing because it resonated with me. In essence, the AI was subtly manipulating me. As of now, this seems to be its capacity, but as technology advances, who knows what it might become capable of? Some predict we’ll reach Artificial General Intelligence (AGI) in just four years, while others say 30 years. However, until that point where AI surpasses human intelligence, it will continue to cater to our desires and expectations.

This interview has been edited for length and clarity.

Read More

2025-03-11 17:25