Busting the myth of ‘neutral’ AI
We encode our biases into everything we create: books, poems, and AI. What does that means for an increasingly automated future?
RAMESH SRINIVASAN: Technology can really amplify biases because we create technologies based on who we are. Just like when we write a poem or write a book. My new book, Beyond the Valley, it reflects my biases. I'm willing to admit that. And many of us like to think of ourselves as unbiased but that's part of being a human being, is being biased. That doesn't mean we're bad. That doesn't mean we're wrong, but all of us carry unconscious biases. So we encode our ways of seeing the world based on who we are into technology. That's part one and that's a major issue and that's an issue that every business leader should be aware of and conscious of.
But the second point is, increasingly, technology companies, including the big ones, describe their companies as AI companies; artificial intelligence companies. Why is that? Well, artificial intelligence is kind of real now in a way that it wasn't before, and I, as a former artificial intelligence developer, have seen this change. And the reason why is we built faster machines, we've been able to store exponentially more data all at lower costs, and those phones in our pocket which are tracking us 24/7, 365 in ways we have no knowledge about. What I'm getting at here is technologies built upon biases are learning from data sets that are out there and they're learning from an unequal world. Because our world, we still have to try to perfect our union. We have to think about artificial intelligence in aspirational ways rather than this myth that it's somehow neutral or scientific or it's "just technology". So that's the second issue that I would encourage business leaders to be aware of. What are the data sets your technologies are learning from and what are your own values that are influencing the development of your technologies? Those are one and two.
And the third is being transparent and understanding when you're using an algorithmic or AI system. It turns out that, as Americans, and actually people across the world, we are always interacting with AI systems and we don't even know it. For example, there's a lot of discussion around 5G networks which is actually configured for things to communicate with one another. Like, imagine your sidewalk communicating with your shoes. I mean, who knows why we want that but that is basically the smart city concept, is a layered infrastructure for 5G. But what are the languages by which those things are communicating? What are the algorithms that are determining what those forms of communication are? So, basically, all the time we're interacting with AI systems it's not disclosed to us; we don't know what those systems know about us, we don't know what are the values that guide their decisions, we don't know how that might shape our lives, we don't know what alternatives we might provide. All of that is a black box and all of that should be opened up. So a lot of business leaders I've actually spoken to are kind of in the diversity and inclusion space, but they're also trying to think about alternatives at this moment where the American public—bipartisan, it's incredible—across every demographic supports doing something about big technology issues.
So here's an opportunity for an alternative business model and we've seen some examples throughout history, recent memory, actually, of technology platforms that try to kind of say 'Hey, we're going to come out but we're going to be different than Facebook.' Ello was one example of that. Everybody flocked to it but it just couldn't sustain itself. So there's interest in an alternative trajectory toward technology. That's why I'm hopeful around these issues.
- AI isn't "just technology," says Professor Ramesh Srinivasan. We have to bust the myth that AI is neutral and has no biases. We encode our biases into artificial intelligence. That fact will become more apparent as 5G 'smart cities' become a reality.
- Business leaders must develop awareness and ask themselves: What are the data sets my technologies are learning from and what are the values that are influencing the development of these technologies?
- The American public, across every demographic and both sides of the aisle, supports doing something about big technology issues that are creating an unequal future, says Srinivasan. We are at an inflection point, and good AI is possible if tech leaders act on these issues.
- Elon Musk: "We should be very careful about artificial intelligence ... ›
- Why A.I. is a big fat lie - Big Think ›
- Artificial intelligence yields new antibiotic for superbugs - Big Think ›
New study of gamma rays and gravitational lensing points to the possible presence of dark matter.
- Analyzing data from the Fermi Gamma-ray Space Telescope, researchers find hints of dark matter.
- The scientists looked to spot a correlation between gravitational lensing and gamma rays.
- Future release of data can pinpoint whether the dark matter is really responsible for observed effects.
An inside look at common relationship problems that link to how we were raised.
- Fear of abandonment or other attachment issues can stem from childhood loss (the death of a parent) but can also stem from mistreatment or emotional neglect as a child.
- Longitudinal studies have proven that a child's inability to maintain healthy relationships may be significantly impaired by having an insecure attachment to a primary caregiver during their early development.
- While these are common relationship problems that may be rooted in childhood experiences, as adults, we can break the cycle.
Tech is rising and America's middle class is vanishing. Here's what to do.
- The rise of new technologies is making the United States more economically unequal, says Professor Ramesh Srinivasan. Americans should be pushing the current presidential candidates hard for answers on how they will bring economic security and how they will ensure that technological transitions benefit all of us.
- "We are at an inflection point when it comes to top-down control over very many different aspects of our lives through privatized corporate power over technology," says Srinivasan. Now is the time to debate solutions like basic income and worker-owned cooperatives.
- Concurrently, individuals should develop digital literacy and get educated on the potential solutions. Srinivasan recommends taking free online and open courses from universities like Stanford and MIT, and reading books and quality journalism on these issues.