谷歌(Google)的人工智能聊天机器人Bard搜索工具,在发布会推介广告中只是犯了一个错误,就让谷歌母公司Alphabet的股价一度大跌7.4%,或相等于市值蒸发了1000亿美元。可是谷歌在网络搜索业务上的主要竞争对手微软(Microsoft)所将引进使用的ChatGPT早已犯下了数次难以被估算的离谱错误,包括向用户承认2+5不等于7等等,但微软的股价却坐地扬升。为何两者会面对截然不同的处境呢?
这或许是因为由微软投资的OpenAI所推出的ChatGPT是以Beta version的形式登场,一开始就郑重声明系统仍不完善,还在供大众测试并提供用户回馈的改进阶段,遂降低了市场对产品的高标准期许。另外,微软的Bing在搜索市场的占有率远远被谷歌抛在后头,市场一开始就以打了折扣的标准,来检视微软Bing搜索引擎将逐步引进ChatGPT来增强用户使用体验的宣布。
反观,谷歌以强大的incumbent姿态,被动的“拉牛上树”推介Bard,内部却可能仍旧面对大规模无预警裁员后所带来低士气、高不满氛围。整个形势可说是恰巧的形成一个完美风暴,来促成谷歌的Bard聊天机器人在发布会演示出错的意外结局。
无论如何,不论是谷歌的Bard亦或是OpenAI的ChatGPT 出错,都可以说是非战之罪。这是因为Generative AI based ChatBot 和Natural Language Processing 其实只是不懂得语义的高效words sequencing algorithm 。
注:
ChatGPT的用户界面有以下声明:
- ChatGPT Jan 30 Version. Free Research Preview. Our goal is to make AI systems more natural and safe to interact with. Your feedback will help us improve
- May occasionally generate incorrect information
- May occasionally produce harmful instructions or biased content
- Limited knowledge of world and events after 2021
In contrast, Google's primary competitor in the search industry, Microsoft, has experienced a rise in its stock price despite its indirectly funded ChatGPT making multiple mistakes, including acknowledging to users that 2+5 does not equal 7. Why do these two companies face such different consequences?
The difference in consequences can be attributed to the way the two companies launched their AI chatbots. ChatGPT was launched as a beta version and was clear from the outset that it was still in the testing phase and needed user feedback for improvement, thereby lowering market expectations. Additionally, Bing, Microsoft's search engine, has a much smaller share of the global search market compared to Google, which means the market applied a lower standard when evaluating Microsoft's incorporation of ChatGPT into its Bing search engine.
On the other hand, Google, being the dominant player in the search engine industry, may still be grappling with low morale and widespread dissatisfaction among its employees in the wake of a sudden and massive layoff. The confluence of these events can be considered a perfect storm that culminated in the inaccurate answer produced by Google's Bard chatbot during the product demo event.
Regardless of whether it is Google's Bard or OpenAI's ChatGPT that makes mistakes, these errors should be viewed as innocent blunders. This is because Generative AI-based chatbots and Natural Language Processing are simply algorithms for arranging words, or in other words, efficient words sequencing programs that lack an understanding of the underlying meaning of language.
Note:
ChatGPT's user interface has the following disclaimer statements:
- ChatGPT Jan 30 Version. Free Research Preview. Our goal is to make AI systems more natural and safe to interact with. Your feedback will help us improve
- May occasionally generate incorrect information
- May occasionally produce harmful instructions or biased content
- Limited knowledge of world and events after 2021
资料来源: https://zinggadget.com/2023/02/09/google-%E8%81%8A%E5%A4%A9%E6%9C%BA%E5%99%A8%E4%BA%BA-bard/
资料来源: https://twitter.com/justthore/status/1618192194955014144
https://timesofindia.indiatimes.com/viral-news/my-wife-says-258-chatgpt-is-being-tricked-with-weird-questions-tweets-viral/articleshow/97794111.cms
资料来源: https://chat.openai.com/chat
资料来源: https://www.reuters.com/technology/google-ai-chatbot-bard-offers-inaccurate-information-company-ad-2023-02-08/
资料来源: https://nypost.com/2023/02/08/google-ai-chatbot-bard-flunks-during-firms-highly-touted-rollout/
0 Comments