4 min read


Using AI to build a better human

More often than not, the discourse about artificial intelligence verses about either the potential misuse or dangers of the technology, racial data bias, bad use of facial recognition applications by governments and corporations or the negative consequences of its deployment on labor and society at large (job losses, humans becoming data farms, robotized societies). But what about using AI to build a better human?

This goes from AI applied to translation or voice recognition, to higher manufacturing automation, to facial recognition or even to analysis and behavioral prediction in the legaltech, fintech, medtech and other areas. The truth is AI can aid humans to become a better version of themselves. However, if old habits die hard, old assumptions die harder: changing our behavior and creating new habits is not easy - however clear the benefits may be. Freeing ourselves from sci-fi assumptions is not any easier. The truth is we need AI to help us become healthier, more productive humans capable of making better decisions with more information and more data. Most true AI offers feedback loops to customize use.

Take the smartwatches and health apps offering personalized versions in a mixture of cloud computing and local data and use. They are only a step further from recommending health-related decisions to help you avoid illnesses. We ought not think of AI in a dichotomy, making a pure choice as something we can or we cannot, we should or should not trust. We should consider AI and its possibilities as an enhancing tool we’re in charge of.

Focusing only on the 'dangers' or 'benefits', on the good or bad simply puts us on a passive mindset, as if we could not be in control of our own actions - whereas this new wave of technology and the future quantum computing are going to expand and multiply our capabilities as humans. I was lucky to be one of the speakers at the recent AI Show in Madrid where I attended a seminal presentation by Raúl Suárez Banegas, Head of Corporate Business in LinkedIn. Raúl spoke about the perfect storm and 3 trends in the new way is seen by employees and employers:

  • 46% of jobs in Germany, France, Italy, Spain and UK are susceptible to automation with technologies that already exist today;
  • 2. 65% of the job descriptions for which we will employ personnel in the generation do not exist nowadays;
  • 3. 26% of the workforce in Western Europe (EU-15 countries, that is 90 million people) are self-employed.

LinkedIn is applying AI to the way the classify and offer results. Their objective is to make “Talent Intelligence” the new era in people’s careers. Talent Intelligence is the use of data and reports with artificial intelligence to make talent the competitive advantage of a company or business. “The best companies will use artificial intelligence applied to data to offer jobs to the best professionals”, he said. LinkedIn uses a whole network of users’ data, company data, job descriptions, skills, schools and educational background, as well as news to compute candidates and options. Gender and race diversity, job offers for people with disabilities, global focus, comparing your company with other similar companies, it all computes in LinkedIn’s algorithms.

His sense of optimism struck me, as well as the clear way in which AI can be put to good use to improve our lives and careers. Raúl’s message is that technology is fine as a complement, but not as a replacement of humans. "The desire to eliminate human labor always generates more work for humans," says anthropologist Mary Gray in her book Ghost Work: How to Stop Silicon Valley from Building a New Global Underclass, referring to what she calls the 'last mile automation paradox'. LinkedIn will propose a list of world-class candidates and save hours of manual selection to HR specialists, but the ultimate decision will be made by human specialists, considering human factors. I cannot imagine a candidate being offered a job without an interview and beating other well-qualified candidates.

In health, researcher Eric Topol, professor, director and founder of the Scripps Research Translational Institute, states in an article in Nature Medicine that almost all physicians dedicated to the practice will use AI in the future. However, he warns that while the field is certainly promising, there is relatively little evidence of such promises. He also warns that the risk of finding defective algorithms is exponentially higher than that of a single doctor-patient interaction, but the reward in reducing errors, inefficiencies, and costs is considerable. AI can be a bridge between big problems and solutions. One of the bridges can be much expanded access to health care across the world. Personally, I’m excited about the bridge to healthier behaviors (exercise, eating habits, weight, etc.) that would obliterate the need for medical treatment in the first place. Throughout human history, humans died from epidemics, wounds and infectious diseases. Nowadays, it’s our behaviors that are killing us. But we can take microsteps to build healthier habits and prevent the skyrocketing rise of chronic illnesses like diabetes or heart disease. This is probably the reason why it makes sense for Google, the data-hungry company to buy Fitbit and gather even more personal analysis and recommendations. I expect the company to grow from exercise monitoring to full health monitoring and medical advice soon.

The end of human doctors?

Internet did not kill doctors’ advice, but made a lot of advice (good and bad) available on the Internet. "Historically, we have always given machines an accuracy value. Most people think that the machine is right, but the truth is that GPS has caused many bad trips and a typo can send us to a completely different place ", says Amando Estela, our CTO.

“The bulk of AI is probabilistic. The key is not only in the science behind it, but how I build that machine: how I curated information, how I got algorithms and neural networks to keep a proper balance in decision-making… but eventually training data offers a result for us humans to apply our common sense and specialist knowledge”, Amando says. “Machines have no common sense, no ability to interpret, no complex communication, no expert thinking. Common sense and general wisdom (what we might call human-quality judgment) are probably the most difficult objectives of artificial intelligence” continues PangeaMT’s CEO Manuel Herranz. “There are people who believe that we will be able to create ultraintelligent machines capable of surpassing any human in all intellectual activities.

We humans can take off-topic information and weigh it accordingly because of personal or professional experience. AI can mimic our decisions, but not all our decisions. What is true is that AI can aid humans to a more human version, to help them make better and more informed decisions about their lives”. LinkedIn is doing precisely that. Our behaviors are already shaped by AI every day. But can we not use AI to pull on those same psychological levers that addict us to games or social media to empower us to disconnect from our tech temptations and have more time and space to connect with what we value most? And what if we did the same around what we eat, how much exercise we do and how much sleep? That’s the way to use AI to augment our humanity, not consume it.


Related Posts

The Creation of Custom Data Sets to Meet Customer Needs: A BSC Project

Rapidly advancing technology and the growing need for accurate and efficient data analysis have led organizations to seek customized data sets tailored to their specific needs. 

Read more

Exploring the Differences Between Human Translation and Machine Translation

The technological advances that have occurred over the course of the last few decades have made it possible to optimize and streamline the work of human translators. One of these advances is machine translation (MT).

Read more

Synthetic Data vs Anonymized Data

What is synthetic data? 

Synthetic data is data that has been artificially generated from a model trained to reproduce the characteristics and structure of the original data. The goal is for the synthetic data to be sufficiently similar to the...

Read more