The staggering speed of improvements in machine learning-powered programs raises many big questions, and the public is only now beginning to discuss them, albeit belatedly. Would it be wise to build machines that one day might be smarter than us and then decide to omit their dumb creators? Conversely, as machines get smarter, keep them in what amounts to a bond What kind of safeguards should be built into these programs? What should society do in the face of an incoming tsunami of plausible fakes? How should we prepare for the promise of a dramatic reshaping of the economy?
All these questions are important. Collectively, they could provoke the most important debate in human history. But this takes a long time. In the meantime, individuals will have to figure out how to answer the small questions AI poses, such as “How can I personally protect myself?”
Hence the discussion about safewords. Voice cloning technology has improved to the point that it’s used by predators to trick unwary people into thinking their loved ones are in serious trouble. When we were talking about one, we both were journalists who provided enough audio and video samples to imitate the machine, and realized we needed a way to say it easily. It’s really me.” that is Kidnapped, arrested, stranded in a foreign country without a wallet. (Notably, the last time this actually happened was years ago.)
We probably aren’t the only ones who need to have this discussion, because the internet contains more information than people realize. Most people now publish their videos and audios on their social media her feed and other people’s feeds. And when you get an urgent phone call from a panicked loved one, it’s not the right time to remember the details of your two lives and try to figure out what the clouds don’t.
This is not the only discussion with family members. We also need to consider how to prove malicious fakes. isn’t it You — an audio of you clearly trash-talking to your boss or cheating on your spouse, for example.
Sorry, I don’t have a good answer for this yet, except for: Now more than ever, it is beneficial to establish a reputation for being extremely honest and trustworthy.
Speaking of bosses, unemployment tends to be a bigger risk than impersonation for many. No, I don’t think AI will make humans completely redundant anytime soon. Most likely, as with the early technological revolutions, some jobs will be destroyed but others will not, and over time workers will migrate to areas where humans can maintain a comparative advantage.
But there is a difference between this technological revolution and the earlier ones, which tended to automate routine and brute force tasks. In contrast, large language models and their brethren threaten to automate the more abstract and skilled parts of many jobs, allowing humans to talk to each other and manipulate various objects without breaking them. It leaves more tedious tasks that are still relatively good, such as doing. they. Machine learning models, for example, have been found to be amazingly good at medical diagnosis, and will probably only continue to improve, but no robots have been invented enough to replace home care assistants.
Perhaps more people would prefer diagnosing an illness to changing an adult diaper. The shift from one to the other becomes painful on both a social and personal level.
That said, it’s not all pessimistic. AI may automate the routine parts of many skilled jobs, freeing the rest to focus on the more fun and creative parts. But at the very least, employees should be prepared for their job descriptions to be rewritten and possibly eliminated.
Start by creating a fearless inventory. What part of the job can the machine learning model do immediately, or after a few faster iterations? How fast will your company or its competitors make such changes? Could you survive the change by learning to use ChatGPT and its many brethren to be more productive and differentiate yourself from the outdated herd? What kinds of jobs can be retrained that are unlikely to be outsourced to computers?
I suspect that many of us have avoided these discussions for the same reason that society has lagged behind in deciding how to address the risks of AI. It’s refreshing, but it’s not clear what to do. That is, pretending that the changes that are already happening to us are not happening.
