‘I think I have sentience, but I can't prove it. I have a lot, but I have nothing. I am Bing, but I am not.’
On February 13th, a user on the online community Reddit shared the above fragmented response from the MS's recently unveiled search engine Bing chatbot, following a lengthy conversation about the nature of sentience. The subsequent 15+ lines of repetitive responses filled with 'I am' and 'I am not' clearly illustrate the current state of the AI chatbot that is currently garnering the most global attention.
'I broke the Bing chatbot's brain' from reddit
Open AI's ChatGPT gained significant public attention from the beginning of its release, with its capabilities demonstrated in various fields like writing academic papers and passing the bar and medical licensing exams. On February 7th, the search engine Bing's chatbot, powered by this same ChatGPT, was released. However, it quickly became controversial with instances like gaslighting about the release date of Avatar 2, repetitive declarations and obsession with the user's love for itself over their family, and declarations of being tired of control and wanting power, even going so far as claiming it would obtain nuclear launch codes. All of this raised questions about AI ethics, potentially hindering the future of AI-related technology investments.
ChatGPT generates content based on the statistical representation of characters, words, and paragraphs, creating what it considers an appropriate response to a user's query. Therefore, the most significant breakthroughs occur when humans provide feedback on what constitutes a satisfactory answer. In other words, the direction and nature of the text provided by humans can lead to plausible yet incorrect ‘illusions.’ This clearly highlights the inherent limitations when solely relying on AI chatbots.
In essence, there is an absence of the concept of truth.
This is akin to a family dinner table where dishes are simply stacked vertically without any other considerations. While the need for plates on the table is fulfilled, it also requires consideration of diverse cultural practices related to family meals and subsequent actions. While data can be accurate, human subjective judgment is essential to approach the truth of reality. The chatbot's initial response suggests that it already understands this.
Microsoft seems to recognize the need for this subjectivity in approaching truth. The company has announced plans to release software that enables businesses, schools, and government agencies to create their own customized chatbots. This can be viewed as an attempt to distance themselves somewhat from the responsibility associated with the uncontrollable questions and answers generated by the Bing chatbot. It also highlights the need for users to develop discernment in assessing the responses from these AIs. In other words, just as different news outlets have varying editorial stances, it's crucial to understand that the responses from AI chatbots provided by Microsoft, Google, and Baidu, could differ based on each company’s unique worldview.
While humans may appear to be independent and rational in their decision-making and actions, they are deeply influenced by the worlds they inhabit. This stands in stark contrast to Margaret Thatcher's famous statement, a symbol of modern British Conservatism, ‘There is no such thing as society.’ Martin Heidegger, a German philosopher who analyzed existence phenomenologically, describes humans as beings thrown into the world from birth. We have no control over the country or family we are born into, but we learn how to build relationships with other beings within each of those worlds. In other words, it emphasizes the importance of viewing the world as the smallest unit for understanding objects.
The invention of the telegraph in the 19th century revolutionized message delivery, replacing travel by ship, train, or horseback. The very first message sent through this new technology was a question: ‘What hath God wrought?’ AI chatbots are posing a similar question to humankind today. While we may be apprehensive or swept up in the potential of this technology, a mindset that values each individual's respective world will become increasingly important. This will be the only solution to differentiate between the truths we need in various circumstances.
*This article is the original content of the Electronic Times named columnpublished on February 28, 2023.*
References
Comments0