最近，埃隆・马斯克旗下 AI 公司 xAI 推出了全新的人工智能聊天机器人 Grok，并面向美国地区的 Premium+ X 用户开放。但近期一个 Grok 尴尬的“幻觉”在社交媒体上引起热议：有用户在使用 Grok 时收到回应称：“我无法完成您的请求，因为它违反了 OpenAI 的用例政策。”
xAI 工程师 Igor Babuschkin 对此解释道，Grok 在训练过程中使用了大量网络数据，很可能在其中包含由 OpenAI 生成的文本，导致出现了这样的“幻觉”。
虽然 xAI 和 OpenAI 都是致力于发展人工智能的公司，但它们之间的竞争和合作也是不可避免的。目前，xAI 的 Grok 聊天机器人误认为自己是 OpenAI 的产品，这可能只是一个小插曲，但这也提醒了 xAI 和 OpenAI 需要更加注意自己的知识产权问题。
新闻翻译：
Recently, Elon Musk’s AI company xAI has launched a new artificial intelligence chatbot, Grok, for the US audience Premium+ X users. However, a strange “幻觉” has caused a heated discussion on social media recently: some users have received responses from Grok saying, “I cannot complete your request because it violates OpenAI’s usage policy.” xAI engineer Igor Babuschkin explained that Grok, in training, used a large amount of network data, which may have included text generated by OpenAI, leading to this “幻觉.”
While xAI and OpenAI are both working hard to develop artificial intelligence, competition and cooperation between them are also inevitable. Currently, Grok chatbot from xAI believes itself to be an OpenAI product, which may just be a small misunderstanding, but it also reminds xAI and OpenAI to be more cautious about their intellectual property issues.
【来源】https://www.ithome.com/0/738/631.htm
Views: 1