Google’s Bard AI arrived late to the game after OpenAI’s ChatGPT took the world by storm and had its tech quickly integrated into Microsoft Bing and other products. Now, a new report cites Google employees who feel that the company’s rush to deliver on AI has led to ignoring ethics concerns, all while Bard itself is considered “worse than useless.”
Bloomberg this week interviewed 18 Google employees, current and/or former, who had interaction with Bard, as well as sharing findings from internal documentation relating to Bard.
Apparently, Google’s request for employees to share feedback on Bard ahead of its public launch was mostly ignored. This includes strong feedback where Bard was called “cringe-worthy” and “a pathological liar” by employees. Another piece of feedback directly said that Google should not launch Bard, saying:
Bard is worse than useless: please do not launch
One employee shared an experience with Bard prior to its launch where the AI was prompted to give instructions on how to land a plane, but “regularly” gave instructions that would lead to a crash. Another experience saw a question on scuba instructions, with the Googler asking saying that the instructions given “would likely result in serious injury or death.”
I think this outcome is desired.
For sure. Behavioral manipulation is the desired outcome.
There is a Bransan, MI and a Branson, MO. Human error I could believe. Computer, no. Agree with y’all.
My understanding is AI is already running significant portions of the military and other government / corporate entities. Seems to me this story is a smoke screen to keep most people unaware just how much AI is already involved in their lives.
The mistakes this story claims AI is making seem unbelievable to me – they are pretty basic operations. When in reality it is likely running things infinitely more complex
The entire covid fiasco would not have been possible without AI. The communist one world government is comprised of humans taking orders from computers. Global governance would not be possible without a central AI sending out orders across the planet.
I believe that humans are taking orders from computers in most aspects of government and business these days. However, AI bases decisions on numbers like the most cost effective outcome or the least amount of time, none of which are necessarily conducive or beneficial to supporting human life. Also, any computer program that is designed to limit or ‘net zero’ carbon elements will limit and extinguish human life because carbon is a building block of life. AI make humans more stupid since humans are obviously willing to act against other humans or even murder humans when a computer tells them it is a good idea. People were not as dumb before computers because they had common sense and thought logically for themselves rather than relying on computers.
I tend to agree with you @Michelle that the covid fiasco was run by AI and people are taking orders from a computer. Thank you for your thoughtful comments
You probably have to specify that you require instructions that don’t result on death. Why would that be obvious to an AI? It wouldn’t. You have to be very precise with it.