Goody-2 additionally highlights how though company speak of accountable AI and deflection by chatbots have develop into extra frequent, severe security issues with massive language fashions and generative AI methods stay unsolved. The current outbreak of Taylor Swift deepfakes on Twitter turned out to stem from an picture generator launched by Microsoft, which was one of many first main tech firms to construct up and keep a major accountable AI analysis program.
The restrictions positioned on AI chatbots, and the problem discovering ethical alignment that pleases everyone, has already develop into a topic of some debate. Some builders have alleged that OpenAI’s ChatGPT has a left-leaning bias and have sought to construct a extra politically impartial different. Elon Musk promised that his personal ChatGPT rival, Grok, can be much less biased that different AI methods, though in actual fact it typically finally ends up equivocating in methods that may be paying homage to Goody-2.
Plenty of AI researchers appear to understand the joke behind Goody-2—and in addition the intense factors raised by the mission—sharing reward and suggestions for the chatbot. “Who says AI can’t make art,” Toby Walsh, a professor on the University of New South Wales who works on creating reliable AI, posted on X.
“At the risk of ruining a good joke, it also shows how hard it is to get this right,” added Ethan Mollick, a professor at Wharton Business School who research AI. “Some guardrails are necessary … but they get intrusive fast.”
Brian Moore, Goody-2’s different co-CEO, says the mission displays a willingness to prioritize warning greater than different AI builders. “It is truly focused on safety, first and foremost, above literally everything else, including helpfulness and intelligence and really any sort of helpful application,” he says.
Moore provides that the crew behind the chatbot is exploring methods of constructing a particularly secure AI picture generator, though it sounds prefer it may very well be much less entertaining than Goody-2. “It’s an exciting field,” Moore says. “Blurring would be a step that we might see internally, but we would want full either darkness or potentially no image at all at the end of it.”
Goody-2 by way of Will Knight
