Vice President of Meta: generative artificial intelligence is still stupid. There is no need to worry about risk.

CTOnews.com July 19 news, Meta vice president and chairman of global affairs, former British Deputy Prime Minister Nick Clegg recently in an interview with the BBC, the risk of generative artificial intelligence (generative AI) downplayed, saying it is still very stupid (quite stupid). He believes that the current hype is "somewhat ahead of technology."

CTOnews.com notes that Clegg's comments come as Meta releases its open-source generative AI model Llama 2, which Meta says has 40% more pre-trained content than Llama 1 and is free for research and commercial purposes.

Generative artificial intelligence refers to artificial intelligence technology that can automatically generate text, images, audio and other content according to input data. Large-scale language models (LLMs) are common generative AI models that can generate coherent text or speech output from a given text or speech input, and Llama 2 is one such large language model.

Clegg's assessment that large language models are "stupid" is not entirely unreasonable. At present, these models also have "illusion" problems, that is, they will give some answers that are completely untrue.

The BBC raised another risk about Llama 2 in its report, namely that the model is open source, meaning anyone can edit the code. Clegg dismisses the idea that they might remove protections against harmful content for malicious purposes as an exaggeration, insisting that Llama 2 cannot even generate images, let alone "create biological weapons." He went on to explain that large language models are always open sourced and that Llama 2 is safer than any other model.

Only time will tell whether these models will be used maliciously. One benefit of the open source model, however, is that it democratizes generative AI. Startups can leverage this code and innovate on top of it, competing with big companies like ChatGPT and Google Bard to advance technology.