“Bender, I don't care whether you have money. I love you for your artificial intelligence and sincerity simulator1."
Artificial Intelligence (AI) has been both a popular, ice-breaker topic of discussion and an object of great concern in recent years, with many experts predicting that
we are on the cusp of a technological revolution that will completely transform the way we live and work – others instead are already prepping for a nuclear armageddon.
While AI has the potential to revolutionize industries, from healthcare to supply chains, and from finance to education, there are also those who think that AI could lead to widespread job losses and the eventual downfall of humanity – remember Skynet?
More realistically, while some worry that such tools could totally replace human workers, the current reality is that AI will likely serve to augment and enhance human capabilities, freeing workers to focus on higher valued-added tasks that require creativity and critical thinking.
However, despite these concerns, AI-based tools are becoming increasingly prevalent in our daily lives, with everything from digital assistants to chatbots and predictive analytics now widely available to the public. One example of such a tool is ChatGPT, a functioning AI chatbot that can have different types of conversations with users.
But, while it is true that AI-based tools are still a relatively new technology and that they are constantly improving, they are surprisingly prone to errors, gibberish overproduction and ridiculously blatant made-up information.
Nonetheless, the tech business environment appears to be eager to put money on AI’s table: in addition to a chatbot recently unveiled by Google, many startups focused on AI are raising substantial funds to develop commercial products for a wide array of industries.
What is ChatGPT?
ChatGPT (GPT sta per Generative Pre-trained Transformer) is an is an Artificial Intelligence chatbot developed by OpenAI and launched on November 30, 20222.
It is built on top of a family of Large Language Models (LLM) – computer programs for Natural Language Processing (NLP) that use deep learning and neural networks – and has been fine-tuned using Machine Learning, a subset of AI techniques which use statistical methods to enable machines to improve with experience.
These models are trained on huge amounts of data available online, in order to generate responses to user questions.
But let’s make an experiment: let’s ask ChatGPT for a description of itself in 100 words:
“ChatGPT is a natural language processing (NLP) model developed by OpenAI that is capable of generating human-like responses to various prompts. It is based on the transformer architecture and has been trained on massive amounts of data to understand patterns and relationships in language. ChatGPT has various applications, including chatbots, virtual assistants, and content creation. With the latest version, GPT-3, having 175 billion parameters, it can generate contextually appropriate and coherent responses that are indistinguishable from those of humans. Overall, ChatGPT is an innovative technology that showcases the power of AI in shaping the future of communication and interaction.”
Got it, didn't you?!
esperimento: chiediamo a ChatGPT di descriversi in 100 parole:
Indeed, the potential of ChatGPT – and that of similar, more specialized tools that are yet to come – is, to say the least, huge: following the release of ChatGPT, OpenAI's valuation was estimated at US$29 billion in 20233!
Yeah, but it doesn’t work perfectly!
ChatGPT and similar AI-based tools have come a long way in recent years, but they are still far from perfect. If pressed, Artificial Intelligence chatbots can spout nonsense, cite unexisting studies to corroborate what they are saying, or even declare that they’re in love with you!
Here are some common errors that these tools can make:
- Context misinterpretation: AI models can sometimes struggle to understand the nuances of ordinary language and context. This can lead to errors in interpretation, where the tool misunderstands the intent behind a user's message.
Little empathy: while chatbots and other AI tools can simulate human-like conversation, they are intentionally designed to lack the empathy and emotional intelligence that comes naturally to humans. This could lead to conversations that might feel a bit impersonal.
Inability to handle complex queries: AI-based tools are quite good at handling straightforward queries, but they can indeed struggle with more complex or abstract questions that require reasoning and inference, such as the precise definition of what feeling they are experiencing at any given moment.
Bias: AI models (be they language models or visual models) are only as unbiased as the data they're trained on. If the data contains biases, such as gender or racial biases, these biases can be reflected in the responses generated by the model.
È It's indeed important to remember that AI-based tools are still a relatively new technology, and they are constantly improving. But fixating on the areas where these tools fall short risks missing what’s so amazing about what they get right: there are already some cases in which real estate workers say that an AI-based tool can perform tasks that would otherwise have taken them an hour or more to accomplish, helping them to move on to more important tasks that can only be performed by humans4.
When these tools work, it’s an entirely new way of interacting with information on the internet – and to do one’s job!
Emerging competitors and AI fundraising news
On February 6th, Google unveiled a new chatbot tool dubbed Bard, in an apparent bid to compete with the viral success of Microsoft’s ChatGPT. The announcement comes as Google's search engine is widely thought to be facing its most significant competition risk in years.
Like ChatGPT, Bard is built on a Large Language Model (LLM)5.
Sundar Pichai, CEO of Google and parent company Alphabet, wrote: "Bard seeks to combine the breadth of the world's knowledge with the power, intelligence and creativity of our large language models. [...] It draws on information from the web to provide fresh, high-quality responses"6.
But there are other giants that are moving fast!
Nearly a year after spinning out of tech giant Alphabet, AI and quantum computing startup SandboxAQ finally announced more funding details this month, saying it had raised a $500 million round. SandboxAQ is examining the related effects of both AI and quantum — which is where the company gets “AQ” — to develop commercial products for telecom, financial services, health care, security and other computationally intensive sectors.
San Francisco-based startup, and rival to ChatGPT, Anthropic is the latest AI company to raise big from a tech giant. The AI startup locked up a $300 million round from Google in February. Anthropic’s new round could bring the company’s total valuation to $5 billion, The New York Times reported. The Financial Times first reported Google as the investor.
By the end of January, Microsoft finally confirmed it has agreed to a “multiyear, multibillion-dollar investment” into OpenAI, the startup behind the artificial intelligence tools ChatGPT and DALL-E, for a reported $10 billion7.
It seems that 2023 is shaping up to be an all-out AI war!
Conclusions
AI-based tools are becoming increasingly present in our daily lives, and the potential for their use is significant. ChatGPT has the potential to revolutionize the way we communicate and interact with information on the internet, leading to a probable revolution in the way we work and we think about work.
However, AI-based tools are still prone to errors which can limit their effectiveness in certain scenarios – imagine an AI-based chatbot trying to school kindergarten children!
Nevertheless, AI fundraising news continues to emerge, and new competitors are appearing, demonstrating the exponential growth of this technology in different industries: for instance, AI-based solutions are being developed to automate and optimize supply chain processes, limiting time losses for human workers dedicated to repetitive tasks such back office data entry, and leading to further improvements in efficiency and cost reduction.
Despite the challenges ahead, it's clear that the age of AI is coming – and while there are certainly risks and concerns to be addressed, the apocalypse is still far away.
Notes
1. Futurama, Season 1, Ep: 10
2. See the ChatGPT page on Wikipedia
3. See the article by Business Insider titled would double the startup's valuation to $29 billion
4. See the article by CNN titled Real estate agents say they can’t imagine working without ChatGPT now
5. See the article by CNN titled Google unveils its ChatGPT rival
6. See this blog article
7. See the article by Crunchbase titled February’s 10 Top Rounds: Real Estate And AI Startups Score Big