PAOLO BENANTI wakes early in the monastery in Rome where he lives to begin each day in prayer. So far, so conventional – for a monk. But outside of the monastery, Benanti’s life is less typical. He trained as an engineer and ethicist, and he now spends his time at the nearby Pontifical Gregorian University, where he teaches the ethics of artificial intelligence. He also frequently walks over to visit Pope Francis himself, to whom he has become a trusted adviser on AI.
This has become an important topic for the pontiff because he fears it could adversely affect the most vulnerable members of society, and increase the world’s already stark inequality. To fight against this, Benanti helped draft , a commitment, signed in March 2020 by global tech firms IBM and Microsoft, that human dignity will always be a priority as AI is rolled out worldwide. With an expanded version of the pact now in the works, żěè¶ĚĘÓƵ spoke with Benanti about the role of religions in the control of technology, if an AI apocalypse is on the cards and whether an algorithm could ever replace God.
Advertisement
Charlotte Lytton: Why did you become a monk?
Paolo Benanti: When I finished high school, I went to study engineering at university. But in the last year, before I got my degree, I decided to change my life and I joined the [Franciscan] order. I’m curious about human nature, what it means to be human, and I was looking to understand reality. I realised that engineering was a good way to approximate reality, but I was starting to look for something more. At that point, I met the Franciscan community and got the feeling that this was where I was meant to be. It offered me life in a fraternity, the opportunity to share and meet people, and the ability to study and search for the meaning of things that had long fascinated me.
What is an average day like for you now?
I live in a really old monastery with four other monks. We wake up in the morning, usually early, and pray together for 45 minutes or an hour. Then, at 7:30am, normal life begins. Everyone has their own jobs – mine is to teach at the university and study. We have communal lunch and dinner, and there are five moments of prayer during the day.
When did your interest in AI begin?
In 2009, after I joined the order, I began a seven-year PhD studying technology and ethics – the idea was to bring together these two worlds. When I started, I felt the most promising technological developments in the next decade or so would be in neurotechnology and nanotechnology. But as time went on, I got the deep sense that AI would be the next big thing.
How did you come to be advising Pope Francis, and how much does he know about AI?
It all started when I spoke at a seminar for Roman Curia personnel, the department that helps the pontiff carry out his day-to-day work. I can’t say too much more about the process.
Can the Pope understand every time we talk about machine learning or algorithms? Probably not. But what he understands very well is the condition of humans around the world. If you tell him that the fate of a migrant or asylum seeker will be judged solely by a machine, he understands what that means. He’s much more active on the value side of the discussion, as opposed to technology and complicated maths.
What are your fears about how AI will shape our future?
Maybe some people think it will be like the apocalypse in the Terminator films. But we should actually be scared by another apocalyptic scenario: the end of the middle class. It’s much, much easier to have an AI act as a surrogate for an accountant than for a blue-collar worker like a plumber or a builder. This means we could have a future in which the majority of people will have low-paid jobs, with the rest taken by machines. That inequality could be really dangerous. The gap created between poor and rich could lead to general conflicts, radical societal transformations and other unknown fallouts.
Are you worried about that future coming to pass?
Well, I see there is huge public opposition to that idea. For instance, in 2021, the United Nations Educational, Scientific and Cultural Organization (UNESCO) set out its , signed by its member states. This ensures that regulations will be put in place to create human-centred AI, so that emerging technologies benefit the population. That major bodies are working on regulations like this makes me a little bit confident that we can handle the challenge.
Tell us about the Rome Call and how it came to be.
I’ve been privy to a lot of panels, not only with governments or academics, but also where technology partners were present. I started discussions with lots of people in different companies. I remember sitting in a coffee break at the end of a panel, and I said: “Wouldn’t it be wonderful if we could more effectively realise the common desire to protect humans with an international call to action?”
I started to open a channel with my contacts in that area. IBM and Microsoft were the first involved. We drafted a list of principles that tried to describe how AI could be developed responsibly – in a way that would respect humans. The principles were transparency, inclusion, accountability, impartiality, reliability, security and privacy. These ideas are connected to a new word that the Pope has used a couple of times: algorethics.
A hundred years ago, John Fletcher Moulton, the mathematician, judge and politician, wrote about the “obedience to the unenforceable”. These are the unwritten ethical standards we hold ourselves to, like when the Titanic was sinking and people urged women and children to get on boats first. Such ethics aren’t written down anywhere and it is not easy to turn these rules into an algorithm. I like to think of these unwritten ethics as the quality of gentleness, and I think machines have to have this quality. That’s why, with the Rome Call, we are trying to encourage companies to shape algorithms to be gentle in the human-machine relationship.
What does the Pope think of the Rome Call?
That it was a really big thing. He is very much behind the Holy See [the government of the Catholic church] being a facilitator for governments, companies and other international institutions to find an agreement on ethical guardrails for AI. For him, it’s crucial that it won’t harm the young generation or the elderly – especially in the wake of the pandemic, in which the importance of our digital lives has exploded.
How can we tell if it is working?
I have to say, something is working. Remember, the Rome Call is just that: it’s a call to action. That means everyone who says it’s important is accountable to themselves to make it workable. This was a cultural moment, and it was a moment in which people with the same sensibility about how dangerous AI could be if not aligned with human values came together. Universities that signed the Rome Call are starting to teach the ethics of AI in their engineering programmes. Journalistic institutions are starting to shape themselves to make the Rome Call effective. Tech companies are developing processes that make algorethics effective in their production processes.
We have a huge event in October with many senior signatories, and in which we will say how algorethics could be included in a lot of new software production processes. How fast and effective it will be depends on how many are involved.
Can we really expect companies to care about ethics over profit?
We cannot ask companies not to think about business. But I have seen in the past five years that – slowly, slowly – they are starting to think about their purpose. Some of them are allowing employees to undertake social purpose projects. I know some who allow 10 to 15 per cent of employees’ working hours to be used for this kind of thing, like digital tools to recognise cyberbullying or pictures of abused children in pornography. Companies are moving gradually from being shareholders to stakeholders. From my perspective, it’s a good sign.
There has been talk of a version of the Rome Call backed by other faiths too. When will that come into effect?
We intended to sign the call in May in Abu Dhabi, with Muslim and Jewish signatories. But the president of the United Arab Emirates, Sheikh Khalifa bin Zayed Al Nahyan, died in May. To allow the nation a period of mourning, we have had to delay. The agreement is ready, we now need to find a workable date for the signing. It’s not easy, with three different religions in which everyone has their own calendar.
Governments are struggling to regulate technology. Do religious institutions stand a better chance?
The majority of people in the world belong to monotheistic religions, so the religious perspective is powerful. The role for religious organisations is to help guide governments. Religions gather human wisdom and timeless values, which, if shared in a non-religious setting, can facilitate dialogue and social awareness. This is what the Rome Call is trying to do, to capture the public opinion that technology should be used ethically.
How might AI be used for good?
Some of my PhD students are looking at this now, researching how healthcare could be transformed by the digital revolution and help to cure disease. Digital tools like AI and the cloud could be used for the global good of humanity. We were able to get vaccinations early and effectively thanks to the compilation of data shared on the cloud. My huge hope is that humankind can find new ways to protect other humans and to use AI for the global good.
A Google engineer recently described the company’s chatbot, which he had been testing, as sentient. What did you make of that?
LaMDA 2.0 was a product designed to guess what the user wanted to hear. This powerful algorithm was able to see that the man sitting in front of it desired to find someone behind the screen. And so what he saw was a mirage in the desert – the powerful tools we have developed fooled one of our technicians with the quality of its answers. Academics are absolutely sure there is no sentient AI.
Can an algorithm ever replace God?
That’s delusional. If God is the source of life, the one who can give us the meaning of life, that is something that is beyond us. If we start to give undue praise to something below us or made by us, then, in the biblical tradition, that has a really clear name: an idol.
Visit the Future Stage at żěè¶ĚĘÓƵ Live to learn more about the innovations shaping the world of tomorrow
