The introduction and development of conversational chatbot technology have gained widespread fame recently. However, the shiny new Chat Generative Pretrained Transformer, a.k.a ChatGPT released by OpenAI in November 2022, seemed to have revolutionized the notion of conversational chatbots.
Ever since its release, the leaders in the tech industry have been struggling to either develop their own products or incorporate the attractive ChatGPT. Microsoft appears to have taken the lead by partnering with OpenAI and taking enormous advantage of the partnership.
Google, however, has been having hard times since its long-awaited Bard botched up during the demonstration. The blunders it made, along with a much-deferred entry undoubtedly put Google on the back foot. Moreover, many employees perceived its attempt to launch Bard as rushed and sloppy. But what’s even more saddening is the fact that Google’s engineers have built something innovative and amazing, only for the executives to kill it. This was the forebearer of ChatGPT.
AI chatbot strangled by chiefs
Google’s chatbot journey dates back a decade ago, in 2013. At the time, the company’s co-founder Larry Page, who was the CEO then, employed Ray Kurzweil. After the hire, Kurzweil started working on several chatbots, one of which was called Danielle. None of them was ever launched.
The AI chatbot that could simulate human talks was begun by Google engineers Daniel De Freitas and Noam Shazeer. The chatbot named Meena was able to discuss philosophical questions and TV shows, as well as to make puns and jokes. The chatbot was created as a side project while De Freitas was working on Google’s YouTube, and was able to mirror human conversations, just like ChatGPT today.
Unfortunately, due to the concerns that it did not adhere to business standards, company leaders prevented it from being tested outside of the company or published as a public demo.
Source: Interesting Engineering
Who was Meena?
Meena, the AI chatbot at the time, was trained on 40 billion words extracted from social media chats. Its main rival OpenAI’s GPT-2, the model that would later become GPT-3, was trained on eight million web pages. Unlike GPT-2, which was made available for testing by researchers, Meena did not receive such approval.
Nonetheless, the work on the project went on and had Noam Shazeer join the team. Shazeer was an engineer at Google’s brain the artificial intelligence research unit. The project got a new name – LaMDA – standing for Language Model for Dialogue Applications. Shazeer added a new type of AI model named Transformer, which enabled programs such ChatGPT to be developed.
Google presented LaMDA to the public, but the chatbot remained in the darkness of the unknown. The duo proceeded to work on the chatbot, finally succeeding in incorporating it into Google Assistant. Again, the public demo was not allowed, so all the tests were performed internally. It goes without saying that the two engineers were frustrated and decided to leave the company in 2021.
ChatGPT as a new screenplay writer
Writing essays, haikus, and songs or poems has become a piece of cake for ChatGPT. It seems to be up to a much greater challenge – screenplays.
The latest South Park episode dubbed Deep Learning mocks the implementation of AI and chatbots in text messages. However, ChatGPT is not only the star of the show, but it also wrote it! Namely, Trey Parker, the creator of the much-loved satirical animated TV series collaborated with ChatGPT to write the screenplay for the 323rd episode that premiered on March 8, 2023.
Judging by the reaction of the audience, ChatGPT was swell of a co-star and co-writer.
But the chatbot doesn’t want to stop here. Once it gets into the world of entertainment, the sky’s the limit. And this is precisely what its president and co-founder Greg Brockman suggests – ChatGPT could assist with improving the “interactive entertainment experience”. He even compared the bot with a team of assistants who might be prone to making mistakes but are enthusiastic and never sleep.
The OpenAI president stated that ChatGPT could help with “drudge work” linked with writing or coding. However, this is not its sole potential. It could also offer a more “interactive entertainment experience.” in other words, you could ask it to rewrite a TV show or film ending, or even make yourself the main star.
You find the Game of Throne’s ending too disturbing? Not a problem, go ahead and write a new one.
How will Hollywood react?
Producers in Hollywood have already begun to make speculations about ChatGPT’s possible effects on the TV and film industries, both favorable and unfavorable. While creative writers may require extra protection from ChatGPT and related technologies, screenplay writers see it as a handy tool. Rather than completely replacing writers, it can facilitate the writing process.
According to Brockman, ChatGPT would be most effective at replacing occupations where consumers “didn’t want human judgment there in the first place,” unlike those that require content control. In addition, Brockman claims that every aspect of life will be, in a way, augmented by this new technology.
However, not everyone is a fan of AI and related AI chatbot tech.
ChatGPT is soulless
Eminent Hollywood director Steven Spielberg and linguist Noam Chomsky have voiced their skepticism about AI amid the ascent of ChatGPT. According to them, AI could significantly affect human’s capacity for autonomous thought and creation. In fact, Spielberg’s pet peeve is handing autonomy over human’s perspective over to the computer. For him, the soul is unimaginable and ethereal, and cannot be created by an algorithm. It simply exists in all human beings.
In line with Spielberg’s viewpoint, Chomsky joined forces with linguistics professor Ian Roberts and director of AI Jeffrey Watumull to co-write an article for New York Times. The trio is of the opinion that ChatGPT, Bard (Google), and Sidney (Microsoft) are all marvels of machine learning (ML). However, chatbots such as ChatGPT “exhibit something like the banality of evil: plagiarism and apathy and obviation.”
With a slight exaggeration, the three men stated that the popularity of chatbots is saddening, considering their amorality, fake science, and inaptitude.
Chomsky further argues that worries about rogue AI may imply that it will never be able to take part in moral discussions or form judgements. In his view, technology might and should only serve as a toy and sporadic tool rather than becoming a significant aspect of our lives.
In defense of itself and its peers, Microsoft Bing’s Sidney has shared its two cents, supporting them with some plausible arguments.
Luckily, though, not all influential figures are so negative about AI and chatbots. Bill Gates, a philanthropist and former CEO of Microsoft, endorses AI evolution. He believes that there is no threat coming from AI models. Even if they do say weird things or go berserk, it’s mainly because users have tricked them into making such mistakes.