In which we review how the search landscape may change with the advent of AI/ML and when can we expect Hybrid Search - Hearch?
Firstly, I do not even remotely believe that the term Artificial Intelligence is even a possibility just yet. The AI we have now is much more accurately phrased as Machine Leaning. I guess that the phrase Machine Learning is not as catchy, cool or futuristic as AI / Artificial Intelligence, hence the general populous buying into AI and not ML. (See What is the difference between ML and AI?)
AI will appear, but only once we have quantum computing working. Progress is being made locally on this at the University of Sussex - see https://www.sussex.ac.uk/news/article/59995-sussex-scientists-make-breakthrough-in-developing-quantum-computers-that-can-solve-big-challenges-of-our-timeWhen it does it will probably casue as many issues than it solves, it has been suggested that quantum computing will be able to crack complex encryption rather easily and that would be bad, very bad!
I digress.
Currently website owners are generally rewarded for good content with good search engine rankings (or SERPS – Search Engine Results PageS) and the current search engines are only being used because they give good (or maybe just satisfactory) results to users for the search terms used.
In fact, we are rewarding Search Engines by using them. Like a cat that purrs or a dog that wags its tail when we stroke them, we reward Google by repeatedly using it because we are happy (or satisfied) with the service it is giving us, and heck we don’t have to pay!
However, we do have some serious emerging alternative services other than the traditional ‘list of relevant things to go and look at’ that are becoming available to us all, and ChatGPT - https://chat.openai.com/ - has taken the world by storm since its public release in late 2022. (See What is ChatGPT?)
The main difference as far as search is concerened is the way ChatGPT presents its results. It doesn't deliver the traditional 'list of things to go and look at’, its methodology is to answer questions asked of it in the form of a conversation. This is very different to what we are used to and it has the ability to completely change the way that we ask for, interpret and use searches on the internet.
It is worth noting that for now ChatGPT has been trained by using data available (think websites, PDF documents, word documents – anything with text) up to the end of 2021. So best not to ask it about the 2022 World cup. (See Who was the best player in the 2022 World Cup?)
So right now ChatGPT is not very good at current affairs, concert tickets or the latest Taylor Swift single (I am more of a Brian Jonestown Massacre man myself…). I am sure that if it was let loose on more up to date data then it would (machine) learn very fast. This does inevitably reduce its usefulness as a ‘Search Engine’ right now, but I am sure that will all change very soon.
How will search change?
I see a scenario where the list of results will be mixed with the natural language results we see when comparing search engines and AI/ML services. I think this will be Hybrid Search or Hearch (you heard it here first :-0). We are going to see Google, Microsoft (Bing) and whoever else start to release hybrid search results that include this mix of results.
Let's compare the results from a simple search phrase “How do I make pancakes?”
If you try this in Google, you get a list of sites that tell you recipes for Pancakes. Try it: https://www.google.com/search?q=how+do+i+make+pancakes
If you try this in ChatGPT you get a completely different response:
“Here's a simple recipe for making pancakes:
Ingredients:
- 1 cup all-purpose flour
- 2 tablespoons sugar
- 2 teaspoons baking powder
- 1/2 teaspoon salt
- 1 cup milk
- 1 large egg
- 2 tablespoons unsalted butter, melted
- 1 teaspoon vanilla extract
Instructions:
- In a large mixing bowl, whisk together the flour, sugar, baking powder, and salt.
- In a separate mixing bowl, whisk together the milk, egg, melted butter, and vanilla extract.
- Pour the wet ingredients into the dry ingredients and stir until just combined (a few lumps are okay).
- Heat a non-stick pan or griddle over medium heat. Use cooking spray or butter to lightly coat the surface.
- Use a 1/4 cup measure to scoop the batter onto the pan. Spread the batter into a circular shape.
- Cook until the edges start to look set and the top is bubbly, about 2-3 minutes.
- Flip the pancake and cook for an additional minute or until the other side is golden brown.
- Repeat with the remaining batter, adding more cooking spray or butter as needed.
Serve with your favourite toppings, such as syrup, butter, fruit, or whipped cream. Enjoy!”
Here’s the thing. I think ChatGPT has answered my question better and has given me the information I am after quicker. It has given me a recipe and cooking instructions. Google just suggested places for me to look - which in the real world my family would probably use a Jamie Oliver or BBC recipe from the list - that is selective search result bias.
Almost unbelievably, Google only had one sponsored AdWords link and that was for a recipe on the RSPCA Assured Egg promotion website, (yes really, here is the URL without tracking code https://www.rspcaassured.org.uk/recipes/traditional-pancakes/.) I was expecting Pancake Mix, Frying Pans and a whole pletera of other adverts…
Hybrid Search will in future, I beleive mix both list and conversational results. You will get the best of both worlds. Maybe you will be asked a retorical question about your question, such as “do you like thin and crispy” or "thick and stodgy", before an answer is generated, this question and retorical-question workflow would be an obvious improvement if it helped get to the best answer quickly.
The Hybrid Search engine will need to improve on traditional results, it will need to be more useful and accurate. If it gives you a crap recipe, you won’t be asking it for recipes again will you? Additionally, it will most likely offer further information, such as links to external resources (maybe even the RSPCA Assured Egg website), alternative recipes, suggestions, links to sellers of supplies and sponsored links (this will be how it is monitized of course.)
The point here is that we are going to see web searching change and my guess is that this will happen VERY soon, Q2 2023 is my prediction, if not Q1.
Should you worry? Well, if you are a search engine user, get excited. It will improve things for us most definitely.
If you are a website owner trying to keep up with all the latest trends to make sure your website ranks organically, then maybe. Maybe because we don't yet know how ChatGPT and other AI/ML services will index and prioritise authorative content yet. But one thing is for sure, if you are creating original well written content that is enjoyed by a readership now you stand a good chance of being noticed and included in the new Hybrid Search.
Remember CONTENT IS KING – and always will be!
Apendix
Conversations with ChatGTP used in this article:
Q to ChatGPT: “What is the difference between ML and AI?”
“Artificial intelligence (AI) is a broad field that encompasses various subfields, including machine learning (ML). AI refers to the ability of computers or machines to perform tasks that would normally require human intelligence, such as visual perception, speech recognition, decision-making, and language translation.
Machine learning, on the other hand, is a specific subfield of AI that deals with the development of algorithms and statistical models that enable computers to "learn" from data, without being explicitly programmed. In other words, machine learning enables computers to improve their performance on a task through experience.
So, while AI refers to a broad concept of machines performing tasks that require human-like intelligence, machine learning is a specific approach to achieving AI through training algorithms with data.”
Well if that isn’t affirmation that AI is actually ML, I don’t know what is.
Q to ChatGPT: “What is ChatGPT?”
“ChatGPT is a conversational AI model developed by OpenAI. It's a language model that has been trained on a large corpus of text data to generate human-like responses to various questions and prompts. It uses the GPT (Generative Pre-trained Transformer) architecture, which is a state-of-the-art neural network architecture for natural language processing tasks.
With its ability to generate human-like text, ChatGPT is widely used for tasks such as question-answering, chatbot development, and text generation. It's designed to be highly flexible and can be fine-tuned for specific use cases, making it a powerful tool for businesses and developers looking to build conversational AI applications.”
Interesting that AI is used in the explanation, but we know AI is now really ML…
Well, jut as a test I did:
Q to ChatGPT: Who was the best player at the 2022 world cup?
The 2022 FIFA World Cup took place after my training data cutoff, so I do not have information on the specific details of the tournament.