AI is ready to solve really big chicken-and-egg problems

As a data scientist-turned-investor who loves to dive deep into technical questions with founders and companies, what excites me most about AI is how it can bring solutions to real-world problems.

While debates swirl around the promise and hype of generative AI fuelled by hyperbolic claims around AI and humanity’s consequences, we might all benefit from shifting the narrative. A focus on the daily, but time- and resource-intensive nature of trivial problems allows us to think about the prospects of applying this new technology in the short term.

We can be sure there will be both mega-winners and big losers. But it is the companies solving practical issues and bringing clear business benefits that are the ones that will survive.

Chicken-and-egg problem

For example, one area where AI really can make a difference is in quality control where humans struggle with attention fatigue. This applies to anything where people have to do the same thing over and over again: anything in which humans have to spot errors and remain consistently attentive. It’s here where we are going to see some of the most interesting industrial AI applications, leveraging human pain points and struggles. 

Another chicken-and-egg problem illustrates my point: AI, wedded to magnetic resonance imaging (MRI) equipment, is being used to determine the sex of unborn chickens by Orbem, a four-year-old Munich-based startup that grew out of the technical university there. 

Applying Orbem technology, a poultry farmer can inspect all their eggs to identify and separate the sexes in their shells. This happens before chicks are born, detecting physical differences between males and females, without penetrating the shell.

Few people know that the chicken-rearing industry slaughters most male chicks because they have no economic value as egg layers themselves. Until now, food processors relied on human inspectors to identify and discard male chicks after they were born. This is not a fun job and it's hugely inefficient for the market.

But when AI is used in this process, it means producers no longer need to incubate and then inspect half of the chicks after birth. The impact is multifaceted - optimising economic value, energy, and carbon footprints all whilst addressing deep-rooted animal welfare issues where one-day-old male chicks would have been culled, in the billions, globally, each year.

Information outweighs capacity

I’ve also spoken with several companies using AI for fire detection in waste and recycling facilities, mitigating improper lithium-ion battery disposal. Far from being a rare problem, there are over 300 fires a year at waste facilities in the UK alone. Utilities can use AI to flag up batteries and other dangerous items in rubbish, notifying workers to step in before they are dumped in crushing machinery and explode.

More broadly, AI’s strong suit lies in sectors that involve a lot of knowledge processing, where the volume of information outweighs the capacity that a single human can process.

Blossom portfolio company AutogenAI, which we invested in in July, uses generative AI to help companies write bids and tenders for public sector and corporate contracts. Bidding for government contracts involves lots of paperwork, as you might imagine, often requiring hundreds of pages of documents to be filed. The founders previously spent 20 years working for leading sub-contractors. They know the pain, firsthand.

As soon as AutogenAI's system recognises that a particular request involves filing similar information as earlier bids or that an answer already exists in a company or public documents, it can automate that process. So Autogen saves time and enables smaller companies to compete alongside bigger contractors. Eventually, that will also save money for governments because it will mean that government suppliers’ costs are lowered. Bidders can focus on the unique intricacies of each tender and the content that matters.

Healthcare and finance

This idea extends to any area where there are huge bodies of data and documentation, which need to be read and processed at a massive scale. Legal work has received a lot of early attention but other sectors such as healthcare and finance are ripe for AI-powered innovation too.

As it is, no lawyer can stay on top of all the changes taking place each year in case law. There’s also vast scope to use AI within e-discovery investigations, where investigators have to extract evidence from millions of emails and might miss key documents. 

In practice, there's no way any human can do these enormous document-processing tasks.

It’s the same with medicine: doctors have too much to do to keep on top of every new research or medical study that gets published. In diagnostics, there are vast areas where AI can help doctors by scanning imaging data to spot anomalies to identify particular issues for which medical professionals can follow up.

Entirely new categories of products are being built with AI. Every company today needs to be thinking about how they could use this technology because, if they don’t, a competitor will be.

This is not about replacing human labour but enabling human ingenuity to accomplish things that just weren't possible before.

Lead image: freepik