Fears are spreading in Silicon Valley that ChatGPT – the AI chatbot taking the world by storm – could become the globe’s go-to search engine.
Google execs are said to have declared a ‘code red’ over fears its $150-billion-a-year search business monopoly could be wiped out thanks to the Microsoft-backed tool.
Much has been made about ChatGPT’s ability to perform eerily-human professional tasks such as writing emails and resumes. But fears in big tech stem from the fact it can instantly conversationally respond to users’ questions, using data aggregated from the internet.
That’s a worry for search engines that rely on users scrolling and researching themselves, exposing them to advertisements. However, experts have noted that ChatGPT appears to have a liberal bias and its answers are not entirely transparent, which raises doubts about its effectiveness as a search tool.
DailyMail.com has compiled an everything-you-need-to-know guide on the AI program:
ChatGPT launched in November 2022 and has since gained in popularity since. The AI-powered chatbot generates emails, naturally converses with customers and provides patients with medical information
What is ChatGPT?
OpenAI invented ChatGPT, a startup co-founded by billionaire and Twitter CEO Elon Musk, and is headquartered in San Francisco.
A system prototype went online on November 30, 2022, and quickly climbed the popularity ladder, garnishing over one million users in its first week.
ChatGPT uses a machine-learning technique called Reinforcement Learning from Human Feedback (RLHF).
This means it learns through interactions with its environment.
The model was trained with 570 GB of data collected from books, web text, Wikipedia articles and other online writings.
It has been fine-tuned using supervised and reinforcement learning – which uses feedback to teach the system what responses are desired.
Proper results draw positive feedback, while incorrect answers or behavior will cause negative feedback.
OpenAI explains on its website that its agent is let loose to act randomly
‘The AI gradually builds a model of the goal of the task by finding the reward function that best explains the human’s judgments. It then uses RL to learn how to achieve that goal,’ according to OpenAI.
‘As its behavior improves, it continues to ask for human feedback on trajectory pairs where it’s most uncertain about which is better and further refines its understanding of the goal.’
The massive dataset of internet text ChatGPT learned from also allows it to generate text similar to what humans would write.
One of the significant advantages of the chatbot is its ability to conversate. It can detect the context of messages and respond accordingly.
This makes it well-suited for language understanding and dialogue generation tasks.
The model can generate high-quality text and human-like responses, making it suitable for a variety of industries, such as customer service, finance and healthcare.
Additionally, ChatGPT can be fine-tuned on a specific task or dataset, making it highly adaptable for different use cases.
This ability provides AI with various applications, such as language translation, text summarization, and question answering.
How does it work, compared to traditional search engines?
The version of the bot available for public testing attempts to understand questions posed by users and responds with in-depth answers resembling human-written text in a conversational format.
Just like Google’s search, it is accessible on a desktop for free, but it responds with information in a totally different way.
Users first have to verify they are not a bot but entering their name and mobile number.
A person can then ask it anything, although some answers will be more detailed than others depending on how much training in that topic the AI has received.
ChatGPT can spit out an answer to many questions in a fraction of second.
The human-like responses to questions could also make it feel like there is another person speaking on the other end.
Experts say a tool like ChatGPT could be used in real-world applications such as digital marketing, online content creation, answering customer service queries or, as some users have found, even to help debug code.
What is ChatGPT’s main purpose?
ChatGPT has been used to pass exams, deliver sermons, write love and apology letters, code software and give relationship advice, to name just a handful of its functions.
Some companies have used the platform to create emails and sales campaigns.
Marketing experts believe its use could be expanded, helping answer frequently asked customer questions and providing base-level customer support.
The healthcare field is excited about its potential too. Some have speculated that it could help answer patient questions about symptoms and advise them on how to continue.
With both the US and UK suffering from doctor shortages, the tool could get patients in need simple care fast, while freeing up physicians for more serious work.
It has also landed a spot as a ‘co-author’ on more than 200 books on Amazon.
Students have also reportedly bent the rules by harnessing the AI’s power to write essays and create projects.
But its ease of use and accessibility has made it a hit with the general public too.
There are anecdotes of people using it to find new recipes or to writing a cover letters for their job search.
How could it be used as a search engine?
ChatGPT differs from other search engines by providing one single answer to a query.
Google, and other similar competitors, spit out hundreds of results and lets users pick their own source of information on a certain topic.
Typical search engines also provide videos, news articles and images to a topic or question.
ChatGPT streamlines this process, though, instead giving users one simple answer based on what the AI itself is trained to say. It also does not show adds.
Google and other search engines like Bing are teaming with advertisements that sometimes portray themselves as actual results.
The AI is trained using text it pulled from the internet. This is unlike search engine algorithms that are instead more personalized, and give results based on a person’s search history.
What Google and ChatGPT output is also different, with the tech giant’s results being based on relevance and popularity.
Chat GPT responses are generated based on probabilities and what it determines to be the context of the conversation.
ChatGPT uses a machine-learning technique called Reinforcement Learning from Human Feedback (RLHF). This means it learns through interactions with its environment, but ChatGPT does not have the ability to think critically
What are its benefits and drawbacks?
Chatbots have been filtering into the internet for decades but have only been able to answer simple questions with standard answers.
But ChatGPT is believed to be the future of these systems, carrying out conversations with users through multiple queries, providing lightning-fast responses and generating software code.
OpenAI’s system can engage with disgruntled customers, increase company sales and provide medical information to patients.
For companies, ChatGPT could help cut costs. But spells bad news for human employees who could lose their position.
Zak Saidi, creative director and AI Lead of creative agency IZSRI, told DailyMail.com that copywriters, bloggers and people who produce online copy could be among the first to fall in the AI revolution.
‘We work with a lot of copywriters and there is certainly a lot of industry talk around the replacement of content writers by AI technology,’ said Saidi.
‘ChatGPT is free, copywriters aren’t – we expect more tech-savvy small businesses to turn to ChatGPT and its AI counterparts when generating more content.’
One of the primary disadvantages of using ChatGPT for customer service is the potential for the chatbot to provide inaccurate or misinformed answers.
Since the GPTs are trained through trial and error, they are only as accurate as the data and algorithms they are based on.
When DailyMail.com asked ChatGPT if it is accurate, it responded: ‘As an AI language model, ChatGPT strives to be as accurate as possible in providing responses based on the information it has been trained on.
‘However, it is important to note that ChatGPT’s responses are generated based on patterns and associations in the data it has been trained on, and may not always be perfect or completely accurate.
‘While ChatGPT can provide helpful information and insights, it is always a good idea to verify information from multiple sources, especially for important or critical decisions.
‘Additionally, ChatGPT may not always have access to the most up-to-date information, as its training data has a cutoff date.’
There are also tasks the AI cannot do, such as brain teasers and giving you a train schedule.
Because the AI only uses information it has been trained with and does not search the internet for it itself, it is not a great tool for current or up-to-date news and information.
There are also concerns over the potential control or information, and the inability for ChatGPT users to choose their own sources.
When a person types a question into Google – or other similar search tools – they are offered hundreds of options to find their answer.
They receive multiple websites, blogs, encyclopedias and other sources that each answer said question in a different way.
While a person will not always choose the perfect source of information, they can pick to find an answer from whatever source they personally deem trustworthy.
ChatGPT is trained using information fed to it by its creators, making it likely the creators’ personal bias affects the types of answers a person receives.
The AI platform was recently under fire for giving answers that supported Democrats, based Republicans, and supported green energy over fossil fuels.
While experts believe ChatGPT will replace Google in as little as two years, the AI-powered chatbot has a different opinion
Another reported issue with ChatGPT and all AI models is their lack of empathy.
‘Chat GPT cannot empathize with users or understand their emotions, making it challenging to provide support in certain situations,’ Paresh M Patel, who has more than a decade of experience in sales and marketing, shared in his blog.
Will ChatGPT replace Google?
In December, Gmail developer Paul Buchheit predicted that ‘AI will eliminate the search engine result page’ and cause ‘total disruption’ for Google.
A New York Times report also said that Google executives sounded a code red within the company amid mounting pressure from ChatGPT.
A core way that Google makes money is from advertisers paying to have their links displayed alongside the results of a search query result in the hope that a user clicks on them.
‘The way I imagine this happening is that the URL/Search bar of the [Google] browser gets replaced with AI that autocompletes my thought/question as I type it while also providing the best answer (which may be a link to a website or product),’ Buchheit said.
‘The old search engine backend will be used by the AI to gather relevant information and links, which will then be summarized for the user,’ Bucheit explained.
‘It’s like asking a professional human researcher to do the work, except the AI will instantly do what would take many minutes for a human.’
The fluency and coherence of the results being generated now has those in Silicon Valley wondering about the future of Google’s monopoly
While some believe ChatGPT will replace Google, the AI has a different opinion.
‘As an AI language model, I do not have the ability to take over any company or organization, including Google,’ it said on the matter.
‘My purpose is to assist and provide helpful responses to users who interact with me.
‘Google is a multinational technology company with a strong market position and a vast array of products and services, so it is highly unlikely that any single entity, including an AI language model like myself, could take over Google.
‘Furthermore, I believe that companies like Google and AI language models like myself can work together to provide even better solutions and services to users around the world.’
Read the full article here