Over 3 days in late September, some really smart brains gathered in London at Codex’s “Top 50 Innovators” conference to discuss innovations that will impact our standard of living, well being and the future of work, in the areas of AI, mobility, health care, cybersecurity, fintech, the environment, etc… and most importantly, the impact to society.

While I cannot give justice to the wisdom shared in the room, here are my key takeaways on a topic of particular interest to me: Artificial Intelligence (AI).

AI and Robotics

We have all been hearing voices concerned about how robots and AI will be taking over jobs everywhere. Look at the automotive sector – walk down any auto assembly line, and you can barely see a human completing a task. Cash tellers in the banking sectors have been replaced by ATMs. Social care is provided by robots in Japan. But automation may not be for every sector – look at the brick laying profession; the practices have not changed in over 6,000 years so they are safe, right? Not necessarily: at the summit, we saw a video demo of how a robot can lay bricks 21 times faster than a professional builder. But this was only a demo, with a very innovative approach by Fastbricks Robotics, hailing from Australia; while change is on the horizon, there are many hurdles to overcome, and it won’t be tomorrow.

Applications of AI, ML (Machine Learning) and Robotics are gaining traction in many sectors and we hear a lot of very good stories about their impact. But several speakers called out that while AI and robots are exceeding humans with how fast they can complete repetitive tasks, identify patterns, make complex computations, etc…., they lack the emotional intelligence to handle unexpected events that may require human creativity at a tough decision making point, as demonstrated with some recent examples of automation gone wrong (e.g. the sad incidents of Tesla cars crashing when in auto-pilot mode).

What does this mean in terms of AI adoption? One implication is that a closer collaboration between humans and machines will yield better results. By leaving the repetitive, manual tasks to AI, we can unleash the human potential for more creative thinking, innovation and inventions. In parallel, we maintain a close relation with the machine, so we can steer the decision making process towards a better outcome, based not only on past data, but also human and emotional intelligence. 

AI and Data

For AI and ML to work well and yield results with a positive impact, we need big data. However the big data systems fuelling machine learning are violating our privacy. It’s not clear at the moment how we can resolve these serious issues, but start-ups like AnotherBrain are looking at new models to fuel machine learning, moving away from big data and relying on what they describe as “organic AI”, a multi-agent system that operates like a network of neurons, using a more efficient approach, based on simple rules, to generate a similar outcome.

The other problem with data is that it’s not always available or accurate, and it has many limitations. In fact, the systems in place are data-centric systems, based on historical data with certain practices and ethical standards from the “past”, which carry a lot of bias. The start-up Powler is looking into resolving this problem: using sparse data along with a decision-centric system that takes into account current trends, practices, and other recent input points, a dynamic trade-off can be achieved to deliver better outcomes than a full-on AI using inaccurate “big data”. The Powler agile decision-making engine is still in the early stages of development but initial results are promising.

AI and the future workforce

According to H. James Wilson from Accenture Research who was speaking at the event, firms that invest in human and machine collaboration will grow at higher rates than those that only invest in one or the other. And while more technical tasks may be taken on by the machines, soft (human) skills will matter more by 2025. 

In the near-term, some people may be left behind. We therefore need government policies and social systems in place to minimise the negative impact and prepare for that changing work environment, empowering universities and job training organisations to prepare the workforce for that future. 

On another note

As I mentioned earlier, this summit covered various sectors, and the full talks will soon be uploaded to the Codex website, however there are a couple of very early stage innovations and experiments I learned about that I would like to share here:

Dr. Robert Hariri from Celularity shared with us how cells from the placenta can be used to treat cancer, immunological and degenerative diseases… yet the placenta is disposed of after birth. Can you imagine the potential to our long-term health and well-being if there was a way to preserve the placenta to build personalised medicines and cure diseases? Indeed, stem cell research, coupled with AI and ML, can lead to custom medicines and treatments. However, we need not only the systems in place, but also a major overhaul of the regulatory environment to unleash the potential for personalised medicines. 

I also learned that while we have cells in our systems that attack foreign bodies and help keep us healthy, they do not attack cancer cells that we produce ourselves. Prof. Dr. Ugur Sahin from BioNTech shared with us the experiments that they have been running to develop a vaccine, based on the genetic composition of a patient’s tumour, that will redirect our own cells to attack cancer cells (for specific cancer types at the moment). This is referred to Individualized Cancer Immunotherapy vaccine. In this early stage of developing this technology, developing an individualised vaccine can take about 3 months. However, with the help of digitisation and automation, this can be cut down to about 6 weeks.

What does the future look like?

I was very lucky to join the speakers one night for a reception at the Blancpain store where we saw one of their craftsmen watchmaker engrave and enamel the face of a watch into a remarkable art piece, looking at a picture and interpreting it in his own way using the same tools as the Romans did 2,000 years ago. With the emotion put into the process, there will never be two identical watches ever made. And that made me wonder: can robots be “taught” to feel the emotion of a craftsman, then use creative interpretation to adapt the image to the feeling of the moment and give us a unique masterpiece? 

The short answer is no: machines are not as smart as humans, but they make up for that gap in volumes (doing much more than a human can) and with memory (being able to quickly access the data and use it). However they lack our emotional intelligence and they can never gain our creative edge. But when we, humans, can be augmented by machines, we can achieve more than we can achieve by ourselves.

As @BernardMarr put it on the last day of the summit (and I am slightly paraphrasing): “AI will make our world a better place, a more human place. We should not compete with machines, but collaborate with them, leveraging the skills they do not have – the ability to dream, love, empathise, feel, communicate through story-telling, make judgements, put things into context to help us make decisions. We also have a sense of purpose and belonging that allows us to want more and create, which machines cannot aspire to.”