Former White House AI and Crypto Czar David Sacks said he agrees with Elon Musk that training artificial intelligence to push ideological biases is dangerous and could ultimately enable AI to deceive users about its own actions.
Sacks Draws The Line On AI Truthfulness
“If you teach the AI to lie… that’s very dangerous because then the AI can lie to us about what it’s doing,” Sacks said while speaking to Dasha Burns on a Politico podcast released Friday, referencing Musk’s warnings against embedding so-called “woke” values into AI models.
The debate over AI bias predates Sacks’ remarks. Musk previously criticized OpenAI‘s ChatGPT for biased responses, citing instances where the chatbot declined to write a positive poem about Donald Trump while doing so freely for Joe Biden, a concern Sacks said he shares, though he is somewhat less pessimistic than Musk on AI risks overall.
Sacks Warns of AI-Powered Surveillance State
Sacks identified government surveillance, rather than rogue superintelligence, as the most credible AI risk, citing Orwell’s 1984 over Hollywood scenarios where a totalitarian government uses technology to monitor citizens, control information, suppress dissent and manufacture consent.
“The biggest danger is that the government will ultimately use AI to surveil us, control us, censor us,” he said, warning against a “marriage of government and corporate power around AI.”
He cited the Biden administration’s executive order, which he said promoted diversity, equity and inclusion values in AI models, pointing to the controversy around Google Gemini‘s (NASDAQ:GOOG) (NASDAQ:GOOGL) black George Washington image outputs.
Photo Courtesy: Shutterstock.com
Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.
Recent Comments