Don’t panic: AI is not taking over the world (just yet)

ChatGPT burst into the spotlight late last year, sparking huge investment but also widespread criticism – Copyright AFP/File Marco BERTORELLO

Deep learning expert Dr John Bates is a “techie original” according to Forbes and Computer Weekly writer Adrian Bridgwater. He is particularly interested in how AI will y come into its own in 2024.

Bates is the author of Thingalytics – Smart Big Data Analytics for the Internet of Things (2015) and a pioneer in the fields of the Internet of Things and Big Data Streaming Analytics event-driven architectures, smart environments and real-time computing, Dr Bates is widely recognized as possessed of deep insight into tech trends.

Bates’ main concern is how politicians are missing the point about AI and deep learning–it’s about written text, not taking over the world and that 2024 will be the most important year for automation of documents since the invention of the printing press, revolutionising enterprises’ comprehension of content and the knowledge they have hidden in all their own content.

In terms of the primary trends expected for 2024, Bates identifies for Digital Journal:

Content sentience

According to Bates: “Over the next year, we can expect to see content develop its own ‘consciousness’. Generative AI tools like ChatGPT have already made it much more intuitive to interact with large libraries of knowledge, as deep learning algorithms distil key points with increasing accuracy and present them back to users in the medium and format of their choosing.”

In terms of the significance, Bates adds: “The next obvious step is for enterprise documents to exist with a built-in understanding of what they are and the type of content they comprise, so they can ‘speak’ directly to recipients or processing teams (or their automated proxies), and even file themselves according to their identified properties, telling relevant IT systems about themselves. They’ll become characters that people and applications interact with and even develop attachments to (so that saving a document to ‘Favorites’ assumes a whole new meaning…).”

This leads onto the year ahead: “2024 will arguably be the most important year for the automation of documents since the invention of the printing press, because of the potential to revolutionize enterprises’ comprehension of all forms of content and the knowledge it contains. Visual walkthroughs of entire company information libraries will become a reality, meanwhile language will become seamless – in that teams will be able to ask English questions of a German document, or the other way around.”

The end of enterprise email

With the next important development, Bates presents: “Enterprise email has been on a sharp downward trajectory for some time now, usurped by more spontaneous and collaborative chat and content exchange platforms like Microsoft Teams.”

Outlining the benefits further, Bates says: “As content becomes more ‘conscious’ and able to relay information about itself, the need for busy and jaded human professionals to have to visually scan, respond/address, or discard individual email messages will cease. Emailed requests, and attached invoices, contracts, or applications will simply announce and identify their existence and file themselves or trigger automated processing, according to their type and priority level.”

Bots will become our proxies at meetings

Bots are set to change the trajectory of the business discussion: “Just as virtual meetings have replaced the real thing, particularly since the first mass-scale lockdowns in 2020, in 2024 we’ll be sending bots to witness what’s being said, take notes, and inject points to consider.”

Explaining further, Bates indicates: “None of this is to say that humans will become redundant. Rather, that tools will only bother people when their expert input or a next-level decision is needed. In the meantime, AI/deep learning tools will do the heavy lifting. There are bound to be some teething troubles, just as early-generation customer service automated loops have driven a lot of people to the edge. But algorithms are improving all the time and as long as teams devote the time to training their bots, the results will be swiftly honed, delighting everyone whose precious time will be saved and work-induced stress reduced.”

This leads to Bates concluding: “Essentially, we’ve entered an era of “institutional memory”, of content collaboration, where AI draws all the findings together and reports the lay of the land, while humans ascend to a higher plane of blue-sky thinking.”

Surface-level RPA will be left in the dust

 On the topic of robotic process automation (RPA), Bates is of the view that RPA “tools are great at doing routine things very well. But, superseded by real, adaptive machine intelligence, RPA as a technology has largely outlived its usefulness. Scraping information from the screen of one app and pasting onto the screen of another means that the technology doesn’t ‘understand’ what it is doing, preventing it from making an intelligent decision about what to do next. To perform a next-level content task requires intelligent insight into the information’s properties and relative importance. In 2024, intelligent content automation will increasingly replace RPA, enabling smarter processing/more advanced automation.”

In terms of the concrete advantages, Bates spells out: “After all, why keep doing mundane things over and over again, when there are ways to run and manage them in more efficient and dynamic ways, aided by AI? Increasingly now we’ll see process automation incorporate deeper learning from content and data. Intelligence which will feed continuously into a growing and ever richer corporate knowledge base.”

The world won’t end because of AI. Not in 2024, at least

Summing up the advantages and disadvantages of AI, together with common perceptions, Bates thinks: “World politicians are missing the point about AI, deep learning, and its potential to do harm. The technology’s great at understanding language, deciphering images, and comparing, cross-analysing and drawing first-line conclusions from the contents of large libraries. It’s also excelling at translating these into easy-to-digest prose and images. But, as far as the technology has come, AI hasn’t yet passed what could be deemed the new Turing test of modern-day machine intelligence – the ability to do the ironing and put clothes away. So we’re not at risk of an AI-driven nuclear war. For now, anyway.”

Bates concludes his 2024 predictive assessment with: “What we can expect is that new AI applications will be ‘crash-tested’ for safety before being unleashed. This is how real-time environments like foreign exchange trading desks will be protected. Currently, around 75% of all AI applications probably risk being stifled by government over-regulation, an anti-innovation stance that must now be challenged if we are to reap the fuller benefits of AI as a society and global economy.”