AI and Blockchain – They Don’t Mix!

AI tests, trials, and learns. The data sets needed for AI are enormous, even for very specific tasks (such as playing Chess).

For AI to be integrated with blockchain would require incredibly fast “transactions”, and huge and ever-growing ledgers of information.

There are some very clear benefits to having AI on the blockchain. Currently, AI operates in a “black box”. Meaning that even AI engineers can’t find out how the answer was reached.

Say we achieve artificial general intelligence (AGI), meaning AI can take on human-like random tasks, rather than being specifically programmed for each.

If this is done under the current model, there is no way to tell if the process leading to the decision is done morally, or within confines that are acceptable.

This is where doomsday theories come from: Without knowing how the decision was reached, and with continuous learning, the AI could learn to manipulate humans to a higher purpose (for the greater good) without us knowing.                    

Welcome to the end of the world.

AI & Blockchain – The Vision

Mixing AI with blockchain could fix this. By having all the data stored on a distributed ledger, open and transparent, the black box issue would be solved.

Of course, this could be done by traditional means, simply by storing all the data.

Doing it on a public, distributed ledger though, would allow anyone to audit the actions. It would give an open playing field, resulting in optimal growth, innovation, and safety.

Sounds amazing, right?

Not so fast.

Why do you think they aren’t storing the data right now?

Because there is a lot of it.

If they were to store the data around AI’s processes, that would be absolutely huge. The data sets AIs learn from are already unfathomably large. Add all considerations and logic the AI runs through, and you end up with colossal amounts of data.

Currently, it’s not just impractical, but nearly impossible to capture all that information.

Will it be possible in the future?

Anything is possible.

Combining ZK technology with AI data storage could have potential. Or any other novel data storage method that lightens loads but keeps the integrity of data.

For crypto and AI to effectively meet, working together for a better future, we first need to solve both technology’s individual problems.

Scalability, scams, and adoption for crypto. Transparency, storage, and preventing the end of the human race for AI.

Only then can we think about combining them, and that is quite far away.

See a case study below, demonstrating what we’re talking about, using the most legitimate and interesting AI crypto project to show the issues highlighted.

To be clear, these problems apply to all crypto AI projects. In fact, the majority are complete scams, or at least pointless. SingularityNET is one of the only legitimate projects, and here you can see why we aren’t buying.

 A Case Study – SingularityNET ($AGIX)

SingularityNET is by far the largest and most respected crypto AI project, focused on creating a network of decentralised AI systems.

“Off-the-shelf AI generally isn’t fit for purpose, and only tech giants can hire teams of engineers to customise the software. SingularityNET is building an automated process that allows any business to connect existing AI tools to build the solution it needs.”

It’s clear to see the goal, but the barriers to overcome are immense.

AI is well known for its comically incorrect answers. We are seeing the first workable model come to market with ChatGPT. Even then, issues arise.

With that in mind, the goal of SingularityNET to connect different AI tools, automatically building out a customised tool which actually works, seems very far away.

Not to mention the fact that these AI tools need to be input to the SingularityNET network, it’s not able to simply take existing tools and combine them.

Then comes the key question, “why crypto? Why does this need to be on the blockchain?”

With enterprise AI in a boom, over the coming years, we will see a huge number of enterprise AI solutions.

These will be customisable, and offer what SingularityNET hopes to offer, but with the benefits of a centralised entity. Someone to blame, customer support, and the ability to be refunded, to name a few. For AI, until it is perfected, these are key!

SingularityNET has a noble dream, to create an open, transparent AI.

Something exciting for sure, but not realistic. At least, not over the next 5 years.

For this product to work, there needs to be the complete success of AI tools in traditional markets, and then a need for that to be moved on-chain.

Cryptonary’s View

Best case scenario: Too early, and in several years, as AI grows, its place in the market will arise.

Base case: Too early, and largely unnecessary. Traditional companies will provide a better means of delivering the service.

Worst case: Completely unnecessary. A noble dream, but nothing more than a dream.