By Guilherme Cintra
Amidst the rapid advances in the field of artificial intelligence (AI), the use (and discussion) of new technology extends beyond laboratories and reaches society as a whole. It is, according to Sam Altman, an “unstoppable revolution.” The CEO of OpenAI, the company behind ChatGPT, visited Brazil last week at the invitation of the Lemann Foundation. During the meeting at the Museum of Tomorrow, he shared his vision of the future of artificial intelligence, its risks, potentialities, and the consequences of the rapid progress of AIs in the world. Below are five significant points from the conversation with Mr. Altman.
1. The speed of change is astonishing – even for AI developers. At the Museum of Tomorrow, Mr. Altman mentioned that they themselves are surprised by the pace of transformations and are striving to keep up with everything, from technical scalability issues to ethical decisions. According to him, the most significant of these was the launch of ChatGPT itself at the end of last year: “We did this knowing that there is a significant cost to accelerating the world. However, I don’t think we can find the answer to the security of artificial intelligence in a vacuum. We truly need interaction and dialogue so that society and technology can co-evolve. The better our systems, the better security research we can conduct.”
Part of the OpenAI team is literally plugging server cables while simultaneously traveling the world to discuss this work. If AI development has an accelerated pace for those on the inside, Mr. Altman also sees the impact on technology users: “I think the biggest challenge is the speed at which society needs to adapt to this. Normally, it takes a few generations to deal with a technological revolution, and now, we might have ten years.”
2. Expanding the discussion is the way to maximize the positive effects of AIs and minimize the negatives. That’s why the CEO of OpenAI is traveling the world. “We want people to have time to collectively figure out how they want this technology to work,” he said. He came to Brazil at the invitation of the Lemann Foundation, which also believes in comprehensive dialogue as a means to mitigate risks and guide the ethical and sustainable development of technology. Mr. Altman fielded questions from the audience, addressing issues from the future of the job market to the regulation of AIs (he advocates an international regulatory model).
Despite the difficulty of keeping up with the speed of these changes, Mr. Altman added: “It’s vital for the world to gradually understand these technologies, to adapt, to understand what’s coming, and for our policymakers to debate.”
3. Ways to “collect value systems” will have to be found to avoid negative biases and ensure that societies with different values and cultures share their worldviews with OpenAI so they can be incorporated. “We can bring humanity’s value system into this technology in a way that we couldn’t before,” Mr. Altman said. “Different people in different places need to be able to push models to align with the values of their countries. I’m optimistic that these systems will be used to reduce the amount of prejudice in the world.”
4. In Sam Altman’s view, the so-called Artificial General Intelligence (AGI) aims to accelerate scientific progress. At the Museum of Tomorrow, the CEO of OpenAI made this explicit several times. For him, Artificial General Intelligence, still very challenging to conceptualize, will allow humans to greatly accelerate scientific advances. It’s about leveraging progress, not replacing people. “What we would’ve done in generations, we will achieve in years,” said Mr. Altman.
It’s worth mentioning that the concept of Artificial General Intelligence is not well defined. In fact, the concept of “intelligence” (without the term “artificial”) is not a consensus. The field of artificial intelligence emerged as a way to try to emulate human cognitive processes through machines. There are specific AIs (for example, ChatGPT, or an algorithm created from data solely to predict school dropout rates) and General Artificial Intelligence, a concept many understand as aspirational. In other words, we aspire for them to genuinely understand things and contexts, adapt to them, and generate any knowledge we seek. Some argue that AGI could reach the level of understanding that a human would—so much so that it could replace and surpass them. However, what Sam Altman tells us is very different from that: “We will still be very useful in a world with AGIs. We will have general artificial intelligence, and humans will still be in complete control of the world.”
5. People will remain at the center. If one of the major concerns about the leap in AIs is related to the future of work, Mr. Altman argued that even the most advanced AGIs cannot replace humans. But he does admit that many jobs will disappear. “Broadly speaking, I think we will have highly adept tools to perform tasks. They are not yet good at completing entire jobs, but they will continue to improve. And we want to be honest about this, even if it’s uncomfortable: job loss happens with every technological revolution.”
Contrarily, an expected impact in the field of work is increased productivity, with professionals accompanied by super-assistants. One example Mr. Altman mentioned is personalized artificial tutors. They will accompany students and understand their specific challenges: “It will greatly benefit education. It will allow teachers to really do what they do best and provide support to students, but they will have a great lever in the classroom, and the results will be astounding.”
There’s one task that machines cannot fulfill, as Felipe Petroski Such, the Brazilian programmer on the OpenAI team, pointed out: defining what humans genuinely want. “I highly doubt that an AGI will be able to figure out what I truly need today to solve my problem. How to know what humans need is something humans still need to figure out,” said Mr. Such.
Lastly, Mr. Altman positioned AIs as tools that will empower humans: “We want to ensure that people have the power to bring things to life without requiring a whole lot of [prior] technical knowledge.”