Recommended for humanity
Nathan Piette Staff Writer I have collected many quotes from Sam Altman. He leads OpenAI, and he leads me to feel many feelings. “I think AI will probably, like - most likely - sort of lead to the end of the world, but in the meantime, there will be great companies created with serious machine learning,” said Altman at Airbnb’s Open Air conference in 2015. Last semester, I made a presentation about organizational changes at OpenAI titled “A Critical Analysis of OpenAI in the Context of Organizational Behavior.” Reviewing it in preparation for this opinion piece meant revisiting leadership conflicts, questions about accountability, and some unsettling statements from Altman. Just then, with my brow furrowed and my hands to the sky, YouTube’s algorithm recommended an interview about how artificial intelligence might actually help humanity. The interview was between Jon Stewart and journalist Josh Tyrangiel, author of “AI for Good.” I expected another conversation about whether computers would save or kill us. Instead, I found a reason to take the space between those possibilities seriously. Tyrangiel described receiving this advice while reporting - “You need to imagine the tech without the tech companies.” That distinction stayed with me. I can question the people developing AI while recognizing that the technology could do extraordinary things, even superhuman things. Supporting medical research does not have to mean giving a handful of billionaires custody of humanity. Imagine a future where AI helps researchers develop treatments for the world’s most murderous diseases, helps doctors recognize problems earlier, and makes medical expertise more accessible. Imagine productivity gains large enough to support short work weeks and universal benefits. There is an even more ambitious and desirable possibility beyond that horizon. There is a society where survival requires less struggle, disability creates fewer barriers, and people have more freedom to learn, create, care for others, and participate in their communities exactly as they see best. Call it utopian. Call it a big Montessori school in the sky. But if there is a chance at extracting that reality from a big data cloud, then we should try to make that nurturing and forgiving world real. We can only get there by cooperating to determine the hurdles to our success, and planning exactly how we get over them. A treatment can exist without it being affordable. An economy can become wealthier without the workers growing with it. AI can help make universal health care and broadly shared prosperity more achievable, but neither arrives without someone in charge of the next software update. Someone decides who gets access, who receives profits, and whose interests matter. That is where my optimism is met with uncomfortable questions. How much privacy are we expected to surrender? Who controls the information these systems collect? What happens to national sovereignty when essential services depend on technology controlled by technology, which is further controlled by a few gatekeeping billionaires? Then there is security, human control, and confidence in our own judgement. Convincing words, images, and voices can be manufactured. We’re making it harder to know what information deserves our trust. Whatever we make of Altman’s apocalyptic remark in 2015, our optimism can’t cloud the catastrophic risk that may be imminent. Fear alone will ensure a miserable future, so I propose finding a slightly more detailed plan. The dystopia is in the details. Tyrangiel’s argument gives optimism a constructive voice. He suggests we have a plan that includes identifying the applications we want, and insists people have a say in how they develop. That means supporting tools to help doctors, teachers, researchers, and people with disabilities. It means demanding privacy protections, meaningful accountability, and economic policies that let the public share in the benefits. We also need elected officials who understand these systems, and they need to prioritize the workers’ voices before they allow the AI market to permanently change the lives of the working class. Let’s all read as much as we can about AI. Let’s read about the record of the midterm candidates on AI. Let’s use our voice now, before it is too late. My YouTube recommendation didn’t resolve the debate. It reminded me that being critical and hopeful can happen in the same conversation. For once, my algorithm recommended hope. Now, let’s do more than like and subscribe.
