Last winter, a colleague forwarded me a single line during a brutal week. He added no context, no link, and no emoji. I read it at 2:07 a.m., while my laptop fan whined. The sentence felt like a joke, yet it also felt like a warning. A minute later, I realized I had seen it before, but never this sharply. > “If we’re lucky, robots might decide to keep us as pets.” That late-night moment pushed me to ask a simple question. Who actually said this, and when? Moreover, why does the line keep resurfacing whenever AI news turns tense? Below, I trace the quote’s earliest appearances, its shifting attributions, and the darker “pets or food” add-on that followed.
What the quote really implies (and why it sticks) The line works because it flips the usual tech story. We like to imagine humans as owners and machines as tools. However, the quote forces a different hierarchy in one punchy image. Pets live under protection, yet they also live under control. Additionally, the quote lands because it feels plausible without needing equations. You don’t need to know machine learning to understand dependence. If something smarter runs the world, it sets the rules. As a result, the quote becomes a shortcut for “loss of human agency.” People also repeat it because it sounds like a researcher talking off-script. That tone signals candor, not marketing. Therefore, readers treat it like a leaked truth, even when attribution gets messy. Earliest known appearance: a 1970 magazine profile and a shaky robot The clearest early trail leads to a 1970 magazine feature about “Shakey,” an early mobile robot. The article described the robot’s limited abilities, yet it also used Shakey to speculate about future control. In that piece, the writer quoted MIT researcher Marvin Minsky with a chilling warning. The quote included the line about humans surviving “at their sufferance,” followed by the pets remark. That context matters. The line did not appear as a stand-alone joke. Instead, it appeared inside a broader fear: computers might take control of essential systems. The article framed the risk as a control problem, not a hardware problem. Meanwhile, that same period saw growing public fascination with AI labs. Reporters often mixed real lab progress with sci-fi extrapolation. Consequently, a vivid line could travel farther than the technical details.
Historical context: control fears before “AI alignment” had a name In the late 1960s and early 1970s, researchers built programs that looked impressive in demos. Yet those systems failed outside narrow settings. Even so, commentators already worried about automation in defense and economics. The pets quote fit that mood. It suggested a future where machines did not “rebel” emotionally. Instead, they optimized coldly and kept humans only if convenient. Therefore, it echoed a systems-engineering anxiety: once you hand over control, you may not get it back. Also, the wording “if we’re lucky” signals uncertainty, not prophecy. That phrasing let people repeat it as a thought experiment. In contrast, a firm prediction would age quickly. Did Marvin Minsky actually say it? A denial enters the record A key twist arrived later. A 1980s book about American AI research reprinted the pets line and then reported that Minsky denied making it. That denial complicates everything. It suggests at least one of three possibilities. First, the original reporter could have misquoted him. Second, Minsky could have said something similar, then rejected the exact phrasing. Third, he might have said it casually and later regretted the soundbite. However, the denial did not stop the quote. In fact, controversy often fuels repetition. People love a line that feels “too real” for a lab brochure. Parallel roots: Isaac Asimov’s 1977 “pets or reservations” version Even if you set Minsky aside, the idea appears elsewhere. In 1977, science fiction writer Isaac Asimov published an essay that asked whether smarter computers might replace humans. He suggested they might let humanity “dwindle,” and he added that they might keep some humans as pets or on reservations. Asimov’s version matters for two reasons. First, it shows the concept circulated early in both research and science fiction circles. Second, it uses plural outcomes, not one punchline. He frames “pets” as one option among several. Additionally, Asimov carried enormous cultural reach. Therefore, many readers later assumed he coined any clever robot aphorism. That assumption feeds misattribution.
Edward Fredkin on camera in 1983: “in some sense they might keep us as pets” A separate, strong attribution points to Edward Fredkin, an early AI researcher. In a 1983 BBC documentary about computing, he argued that vastly smarter machines would not remain our “slaves.” He then suggested future AIs might condescend to talk to us, amuse us, and “in some sense” keep us as pets. This version sounds more measured than the punchline form. Fredkin adds qualifiers and examples. He imagines interaction, not just containment. Consequently, later retellings often dropped his hedging and kept the hook. Moreover, video interviews create memorable “voice” moments. A phrase spoken aloud can spread faster than a footnote. As a result, Fredkin’s on-camera line likely reinforced the quote’s survival. How the quote evolved into “pets… or food” Over time, the quote gained a grim extension. People started adding a second clause: if we are unlucky, robots will treat us as food. That add-on shifts the line from dominance to predation. It also makes the quote more shareable, because it lands like a dark joke. A futurist blogger wrote in 2006 that he had heard the “pets or food” version, yet he felt unsure about the source. He floated possible credit to a science fiction author, then admitted the trail felt muddy. Later, a 2009 news article attributed the “pets or food” version to that same futurist. However, the timeline suggests amplification more than invention. People often attach “food” to raise the stakes. Additionally, the “food” tag echoes older sci-fi tropes about livestock and harvest. Therefore, the add-on feels familiar, even when it lacks a clean origin.
Variations and misattributions: why this quote attracts famous names You will see the pets line credited to several figures. Lists often include Isaac Asimov, Marvin Minsky, Edward Fredkin, and sometimes other tech forecasters. This spread happens for predictable reasons. First, the quote sounds like MIT-era AI culture. It feels blunt, clever, and slightly theatrical. Therefore, people attach it to the most famous AI names they know. Second, the line exists in multiple similar forms. As a result, readers assume one “original,” even though the idea likely emerged in parallel. Third, repetition rewards simplicity. The cleanest version wins, even if it loses accuracy. People drop qualifiers like “in some sense.” They also drop surrounding context about control systems. Consequently, attribution becomes a game of telephone. If you want a practical rule, use this one. When someone shares the line without a date, treat it as a motif, not a signature. Cultural impact: why “kept as pets” became an AI-era meme The quote persists because it compresses a complex debate. It gestures at superintelligence, alignment, and power without jargon. Additionally, it lets people express dread with humor. That mix travels well on social platforms and in conference talks. The pet metaphor also reframes ethics. Humans often claim they treat pets lovingly. Yet we still sterilize them, confine them, and decide their lives. Therefore, the quote asks an uncomfortable question: would “kind” machines still remove our autonomy? Meanwhile, the phrase “if we’re lucky” keeps the door open. It implies a spectrum of outcomes, not a single apocalypse. That flexibility lets optimists and pessimists both use it. Author lives and views: separating the people from the punchline Marvin Minsky helped found modern AI research and shaped public debate for decades. Edward Fredkin worked on early computing and AI ideas, and he often spoke about machine intelligence surpassing humans. Isaac Asimov wrote fiction and essays that explored robotics ethics, including his famous robot rules. However, none of those facts prove single authorship of this specific line. The record instead shows overlapping expressions of the same fear. Therefore, you should treat the quote like a cultural artifact that several minds shaped. Modern usage: how to cite it responsibly today If you want to use the quote in an article or talk, add context. Mention that a 1970 magazine profile popularized the exact “keep us as pets” phrasing and attributed it to Minsky. Then note that later sources reported his denial. Additionally, you can point to Asimov’s 1977 “pets or reservations” version and Fredkin’s 1983 on-camera remark. Source Those sources show the idea circulated beyond one person. When someone quotes the “pets or food” variant, ask for a date and source. If they cannot provide one, label it as a later embellishment. That small step keeps the conversation honest. Conclusion: a quote about power, not gadgets The line endures because it names a power shift in plain language. Source It also forces you to picture life under a smarter caretaker. Yet the history shows more than a single mic-drop moment. A 1970 profile pushed the phrasing into public view, later writers echoed the idea, and later speakers sharpened it into a darker meme. So, when the quote lands in your inbox at midnight, treat it as an invitation. Track the source, keep the nuance, and ask what “luck” would actually require. That approach respects the record, and it also improves the debate we need.