Famed whistleblower Edward Snowden recently spoke virtually at CoinDesk’s Consensus 2023 conference. Snowden expressed his optimism for the future of artificial intelligence, but there is a catch.
“As with children, we don’t need machines to be like us…we need them to be better than us. And if they aren’t better than us, we did a terrible job.”
Snowden, who gained fame from exposing government surveillance, said that sophisticated AI models such as ChatGPT may invoke the government and big tech to target users’ data. To prevent this, he insists that these software developments remain open-source, implying that recent developments are straying from this nature; he specifically called out OpenAI on this topic.
See Related: Whistleblower Edward Snowden Played A Key Role In The Creation Of Zcash
Training AI To Be Better Than Us
Snowden believes AI is being taught to think like us based on how data is collected and filtered for its language processor. Currently, these AI models are fed vast amounts of data and online content from both web pages and social media; he argues this isn’t ideal if the AI is to act smarter than us.
”They’re training [AI models] on Reddit threads…it’s like the internet equivalent of YouTube comments. But you want to create something decent, good, that’s creative and useful.”
See Related: Regulation Hits Generative Artificial Intelligence; Security Concerns And Privacy Cited
A recent study found that a few experts believe that within 45 years, AI will surpass humans “at more or less everything”. For machines to truly be smarter than humans, not just more efficient, they would need to be encoded with things like creativity and emotion – both of which are a big reach with current technology.
Human intelligence is based on various factors growing up, both good and bad; for a machine to be better than a human, the question remains, how would these vital experiences be coded and learnt by a model stuck inside of a computer?
