As the New York Times reported, as the AI race among big tech companies heats up, “worrying about being able to fix it later is absolutely fatal right now,” a Microsoft executive said in an internal email about generative AI. wrote to
In other words, to quote Mark Zuckerberg’s old motto, it’s time to “act fast and break things down.” Of course, if something breaks, you may have to fix it later, but that will cost you money.
In software development, the term “technical debt” refers to the implicit cost of future fixes as a result of choosing a faster, less careful solution now. Sometimes rushing to market means releasing software that isn’t ready, knowing that once it hits the market we’ll know what the bugs are and hopefully fix them.
However, negative news stories about generative AI tend not to be about this kind of bug. Rather, many of the concerns are that AI systems are amplifying harmful biases and stereotypes, and that students are using AI deceptively. We hear about privacy concerns, people being misinformed, labor exploitation, and fears about how quickly human jobs will be replaced, to name a few. These issues are not software bugs. The perception that technology reinforces oppression and prejudice is very different from the perception that website buttons don’t work.
As a tech ethics educator and researcher, I’ve thought about this kind of “bug” a lot. It’s not just technical debt here, it’s ethical debt. Just as technical debt can arise from limited testing during the development process, ethical debt arises from not considering possible adverse effects and social harm. And especially in the case of ethical debt, the person who owes it is unlikely to end up paying it.
off to the race
I imagined the debt ledger would start filling up as soon as OpenAI’s ChatGPT, the starter pistol in today’s AI race, was released in November 2022. Within months, Google and Microsoft released their own generative AI programs, seemingly rushed to market to catch up. Google’s stock price fell after the company’s chatbot Bard confidently provided incorrect answers during the company’s demos.
Some expect Microsoft to be particularly cautious about chatbots, given Tay, the company’s Twitter-based bot, which was shut down almost immediately after making misogynistic and white supremacist remarks in 2016. There may be But his AI-powered early conversations with Bing unsettled some users, repeating known misinformation.
As the social debt of these hasty releases comes due, we will hear mention of unintended or unforeseen consequences. After all, even with ethical guidelines in place, OpenAI, Microsoft, and Google don’t see a future. How can we know what social problems will arise before the technology is fully developed?
At the root of this dilemma is uncertainty. This is a common side effect of many technological innovations, but it is even greater with artificial intelligence. After all, one of the great things about AI is that its behavior is not known in advance. AI may not be designed to produce negative outcomes, but it is designed to produce the unexpected.
However, it would be disingenuous to suggest that engineers cannot infer accurately about many of these results. There have been countless examples of AI reproducing prejudices and exacerbating social inequalities, but these problems are rarely acknowledged publicly by the tech companies themselves.
For example, it was outside researchers who found racial bias in a widely used commercial facial analysis system and in a medical risk prediction algorithm applied to about 200 million Americans. . Academics, advocacy groups, and research organizations such as the Algorithmic Justice League and the Distributed AI Research Institute do much of this work of identifying harm after the fact. And even as companies continue to fire ethicists, this pattern is likely to continue.
Guess – responsibly
I sometimes describe myself as a tech optimist who thinks and prepares like a pessimist. The only way to reduce ethical debt is to take the time to think ahead about what could go wrong, but this is not something engineers are always taught to do.
Scientist and iconic science fiction writer Isaac Asimov once said that science fiction writers “foresee the inevitable, problems and catastrophes may be inevitable, but solutions are.” ‘ said. Of course, sci-fi writers aren’t often tasked with developing these solutions, but engineers who are currently developing AI are.
So how can an AI designer think like a science fiction writer? One of my current research projects is focused on developing methods to support this process of ethical thinking. increase. I don’t mean to design with far-off robot wars in mind. What I mean is the ability to consider future outcomes, including the immediate future.
This is a subject I have explored in my teaching for some time, and I encourage my students to ponder the ethical implications of science fiction technology to prepare them to do the same with the technology they create. One of the exercises I developed is called “The Black Mirror Writer Room” where students speculate about the possible negative impacts of technologies such as social media algorithms and self-driving cars. Often these arguments are based on past patterns and potential villains.
PhD candidate Shamika Klassen and I evaluated this teaching practice in a research study that asked computing students to imagine what might happen in the future and brainstorm ways to avoid that future in the first place. We have found that there are educational benefits to encouraging
However, the aim is not to prepare students for the distant future. It’s about teaching thinking as a skill that can be applied immediately. This skill is especially important for students to imagine harm to others. Harm from technology often disproportionately impacts underrepresented and marginalized groups in the computing profession. The next step in my research is to apply these ethical guessing strategies to real-world technology design teams.
Do you have time to pause?
In March 2023, an open letter with thousands of signatures proposed a moratorium on training AI systems more powerful than GPT-4. If unchecked, AI could “eventually outnumber, become smarter, become obsolete, and be superseded by us,” or even cause a “loss of control over our civilization.” The authors warned that there is even a
As the letter’s critics point out, its focus on hypothetical risks ignores the actual harm that is happening today. Nonetheless, among AI ethicists, the need to slow down the pace of AI development — even if developers raise their hands and cite “unintended consequences” as a reason — slows down AI development. I don’t think there’s much of an objection to the point that it doesn’t.
It’s only been a few months since the “AI race” has accelerated significantly, but I think it’s already clear that ethical considerations are being left behind. But debt will eventually come due, and history suggests that executives and investors at big tech companies may not be the ones to pay it.
