Will AI wipe out humans? Where is the bottom line of artificial intelligence?
Who would poison GPT? A lot of people. When was Dark Web GPT born? Probably already in the works. Is it still too late to put AI in a cage? It's time, but where is the cage?
I studied GPT's armor cages and walls. GPT is like the world's first living dragon, and soon began to face the cruelty of nature after being put into the world. Hackers around the world, adventurers with ulterior motives, anarchist underground organizations, genius lone wolves who don't like to bathe, and tens of millions of bored smart people are wondering if GPT4 is not immune to swarming attacks. All similar products that follow later will have problems.
It seems to us that artificial intelligence is not as hot as in the first two months, but the madness has just begun. This is the most chaotic juncture in the history of technology, and the fear of falling behind is far greater than the fear of artificial intelligence. Let's share the unpredictable large-scale hacker high-stakes contest that is extremely naked without a risky solution.
You seem to be able to see Musk, Wang Xiaochuan, Li Yanhong and others, naked in the wilderness racing to catch up. Open AI AI can kill me, risk can kill me, but opponents can't kill me. The long-term implications of this competition will be discussed later.
Let's first look at the specific urgent security risks, hard risks always make people, and the dangers caused by people run through the whole process, several of which are more representative.
The most intuitive and widely circulated force for data breaches is Samsung. Complained that GPT leaked its confidential data because they had two employees who uploaded sensitive code and asked GPT to help fix bugs. Another employee uploaded the confidential meeting content and asked GPT to help write the conference machine, and passing it up was equivalent to making it public. So now many technology manufacturers have banned AGAG. This makes the programmer very frustrated, in fact, this is fine, it is an honest problem. For deep thinking about artificial intelligence, follow me and watch the next episode.